| *** sabeen3 has quit IRC | 00:01 | |
| *** dmsimard is now known as dmsimard_away | 00:01 | |
| *** afazekas has quit IRC | 00:01 | |
| *** jcoufal has quit IRC | 00:02 | |
| *** markvoelker has joined #openstack-infra | 00:04 | |
| *** luqas__ has joined #openstack-infra | 00:07 | |
| *** markvoelker has quit IRC | 00:09 | |
| *** luqas has quit IRC | 00:09 | |
| greghaynes | mordred: oh? youre wanting to add a vhd output format to DIB? | 00:12 |
|---|---|---|
| greghaynes | mordred: if you know of a good way I can test the resulting image id be up for hacking on that | 00:13 |
| openstackgerrit | Maru Newby proposed openstack-infra/project-config: Add job to run neutron func setup https://review.openstack.org/156421 | 00:13 |
| greghaynes | also, vhd, really? | 00:15 |
| *** baoli has joined #openstack-infra | 00:19 | |
| mordred | greghaynes: I do not know a good way | 00:20 |
| mordred | greghaynes: I know the current way I use :( | 00:20 |
| greghaynes | is that to upload to rax? | 00:20 |
| mordred | yup | 00:20 |
| mordred | qemu-img convert -O raw ubuntu.qcow2 ubuntu.raw | 00:20 |
| mordred | sudo docker run -v $(pwd):/images emonty/vhd-util convert -s 0 -t 1 -i /images/ubuntu.raw -o /images/intermediate.vhd | 00:20 |
| mordred | sudo docker run -v $(pwd):/images emonty/vhd-util convert -s 1 -t 2 -i /images/intermediate.vhd -o /images/ubuntu.vhd | 00:20 |
| mordred | greghaynes: I believe step one is "make vhd-util not come from a docker image" | 00:20 |
| mordred | which likely means an PPA for now | 00:21 |
| mordred | or something | 00:21 |
| greghaynes | aye, is that the deal we docker because its part of xen essentially | 00:21 |
| mordred | well, kinda - it's actually because you need a patched version of vhd-util | 00:22 |
| mordred | and I didn't feel like actually dealing with debs | 00:22 |
| mordred | greghaynes: https://github.com/emonty/vhd-util/blob/master/build.sh | 00:22 |
| *** mjturek1 has quit IRC | 00:22 | |
| openstackgerrit | Maru Newby proposed openstack-infra/project-config: Add job to target only neutron functional setup https://review.openstack.org/156421 | 00:23 |
| mordred | greghaynes: so, I mean, turning that build script into a debian package won't be rocket science | 00:23 |
| greghaynes | heh, yes | 00:23 |
| mordred | greghaynes: but ultimately I'd "love" to do -t qcow2,vhd and just have happiness ensue | 00:23 |
| mordred | because then there would actually only be one other difference in the playbook - and honestly, it's adding two rax-specific parameters that would be no-op on hp anyway | 00:24 |
| greghaynes | yea, and we can just make dib check for $(which vhd-util) for that | 00:24 |
| mordred | yup | 00:24 |
| jogo | tchaypo: yes, but I have a question for you too | 00:26 |
| jogo | tchaypo: it sounds like ttx reached out to stable-maint about this but not sure where that took place | 00:26 |
| *** sarob has joined #openstack-infra | 00:26 | |
| *** Longgeek has joined #openstack-infra | 00:27 | |
| *** dalgaaf has quit IRC | 00:27 | |
| *** dims_ has quit IRC | 00:27 | |
| marun | sdague: I've proposed a new job that just runs the neutron functional job's setup. Given that it won't be subject to the vagaries of test execution, I'm hoping you'll be willing to have devstack gate on it once it proves stable. | 00:28 |
| *** mtanino has quit IRC | 00:28 | |
| tchaypo | jogo: was that your question? I don’t know what happened either | 00:29 |
| *** dims__ has joined #openstack-infra | 00:30 | |
| *** asettle has joined #openstack-infra | 00:30 | |
| tchaypo | except that I’ve been looking for something to do, and mordred mentioned that he’d like it if someone could help out, and I thought it sounded roughly like something I might be interested in, so I’m now poking around to see if it is something I want to pick up | 00:30 |
| tchaypo | I’m only just getting through the thread from a week ago now - I didn’t see it at the time | 00:31 |
| *** jcoufal_ has quit IRC | 00:31 | |
| *** sarob has quit IRC | 00:31 | |
| marun | jogo: the skip-if bug is pretty obvious in retrospect | 00:32 |
| jogo | tchaypo: hmm | 00:32 |
| marun | jogo: it predates skip-if, actually. | 00:32 |
| marun | jogo: every attribute of metajobs is copied only if it's explictly defined...except for voting, which defaults to true | 00:33 |
| jogo | marun: ohhh that is fun | 00:33 |
| marun | jogo: so, I'll have to ensure that jobs default to voting=True but metajobs default to something like None | 00:34 |
| marun | jeblair: so, storyboard for zuul bugs? | 00:34 |
| clarkb | I dont think its just the voting value | 00:35 |
| marun | clarkb: there is one other boolean, yes | 00:35 |
| marun | clarkb: I'd handle both of them | 00:35 |
| clarkb | and yes storyboard | 00:35 |
| marun | clarkb: hold_following_changes | 00:35 |
| jogo | tchaypo: so one place where stable maint is missing tooling is around dependencies | 00:35 |
| jogo | tchaypo: pinning direct stable/juno requirements still leaves space for secondary requirements to break us | 00:36 |
| marun | clarkb: thanks | 00:36 |
| jogo | and the second part thing that is needed is http://lists.openstack.org/pipermail/openstack-dev/2015-February/057089.html | 00:37 |
| marun | clarkb: is linking supported in commit msg like for launchpad? | 00:37 |
| clarkb | marun no not yet, just do it as if it were though | 00:37 |
| clarkb | reviewers know to look it up | 00:37 |
| marun | clarkb: how to tell between lp id and sb id? | 00:37 |
| clarkb | they numbers dont overlap | 00:38 |
| greghaynes | I believe they just carry over | 00:38 |
| marun | clarkb: ah, ok | 00:38 |
| jogo | tchaypo: and probably more transparent reporting of stable branch statuses somehow. So its easier to see if its working not working etc. | 00:38 |
| *** sdake has quit IRC | 00:38 | |
| tchaypo | taking those one at a time | 00:39 |
| tchaypo | the “Dependencies” you’re talking about would be things like | 00:39 |
| tchaypo | oslo-sphinx depends on sphinx, so we cap that; but we don’t directly mention pygments, so we could be broken by a new version of that | 00:39 |
| tchaypo | so the “tooling” in this case would be building an exhausting list of transitive dependencies and pinning them? | 00:40 |
| tchaypo | (which I believe I’ve heard still doesn’t fully give us reproducible environment builds, but it’s a step) | 00:40 |
| jogo | tchaypo: correct | 00:40 |
| jogo | so harlowja_away has played with this a bit | 00:41 |
| jogo | one way to do this though is: take pip freeze from a working job | 00:41 |
| jogo | and re-install via 'pip install --no-deps -r pip-freeze.txt' | 00:41 |
| jogo | or something like that | 00:41 |
| mordred | greghaynes: I just made debian packaging for vhd-util - testing now | 00:42 |
| greghaynes | oh, well then | 00:42 |
| clarkb | jogo maybe first step is fixing your pins :) | 00:42 |
| tchaypo | I could swear I saw a conversation the other day where someone asserted that this isn’t sufficient | 00:42 |
| clarkb | tchaypo its not | 00:43 |
| clarkb | setup.py is turing complete | 00:43 |
| jogo | tchaypo: something like this http://nvie.com/posts/better-package-management/ | 00:43 |
| jogo | clarkb: well they work, just aren't pretty | 00:43 |
| openstackgerrit | greghaynes proposed openstack-infra/storyboard: Use python3 compatible urllib https://review.openstack.org/155150 | 00:43 |
| clarkb | it can do whatever it wants and often do | 00:43 |
| tchaypo | I think the specific thing they were talking about was cases where packages in the system leak into the venv, so installing the same pinned deps on different base systems can lead to different installed code? | 00:43 |
| clarkb | jogo they work for a subset of packages | 00:43 |
| clarkb | that list isnt static | 00:43 |
| jogo | tchaypo: ahh yes that is correct | 00:44 |
| jogo | clarkb: no, I just added caps | 00:44 |
| jogo | you have an example of something that needs fixing? | 00:45 |
| clarkb | jogo >=0.2.0,<=0.2.0 | 00:45 |
| jogo | clarkb: that works :) | 00:45 |
| clarkb | that should be either ==0.2.0 or >=0.2.0,<0.3.0 | 00:45 |
| jogo | but >=0.2.0,<=0.2.0 is still valid and works | 00:46 |
| clarkb | its valid for pip | 00:46 |
| clarkb | will only confuse humans | 00:46 |
| jogo | correct | 00:46 |
| clarkb | yes so should be fixed... | 00:46 |
| * greghaynes has flshbacks of the BASIC notequal operator | 00:47 | |
| clarkb | and ideally bumped to <0.3.0 if project semvers | 00:47 |
| tchaypo | stepping back to "one way to do this though is: take pip freeze from a working job and re-install via 'pip install --no-deps -r pip-freeze.txt’" | 00:47 |
| tchaypo | what is “this” here? | 00:47 |
| clarkb | tchaypo reproduce an installed wnv | 00:47 |
| tchaypo | Is that meant to mean “One way to test if we have a full set of deps”, or was that something else? | 00:47 |
| clarkb | s/wnv/env/ | 00:47 |
| *** bknudson has joined #openstack-infra | 00:48 | |
| tchaypo | okay, I think I understand. | 00:48 |
| jogo | clarkb: well I am not sure I agree. As capping reqs means something different here. it means this is what upstream is testing | 00:48 |
| jogo | not beyond this won't work | 00:48 |
| jogo | anyway we at least need to better clarify this | 00:49 |
| clarkb | jogo no thats pinning | 00:49 |
| clarkb | capping is what I describes | 00:49 |
| clarkb | anyways to avoid confusion we should reduce the pins/caps as possible | 00:49 |
| jogo | tchaypo: yeah --no-deps means you can specify the versions of everything and dependencies aren't installed. But you still need to sort out the system vs pip installed stuff | 00:49 |
| jogo | clarkb: well long term I think we need to pin (==) everything | 00:50 |
| tchaypo | it sounds to me like —no-deps is only useful after you have a fully specified list of requirements | 00:50 |
| jogo | and have a requirements.in and .txt file | 00:50 |
| clarkb | jogo I disagree but ok | 00:50 |
| jogo | tchaypo: yes | 00:50 |
| jogo | clarkb: oh? | 00:50 |
| tchaypo | and getting that fully specificed list is at least half the battle | 00:50 |
| jogo | tchaypo: yes, its not easy | 00:51 |
| tchaypo | although in theory pip-compile should do that for us | 00:51 |
| clarkb | I stopped working on this stuff because no one agrees witg me on how it should be fixed :) | 00:51 |
| jogo | well you have to make sure the full list passes tempest | 00:51 |
| clarkb | so I am letting others do it for the most part | 00:51 |
| jogo | just because we can install everything doesn't mean it won't cause a bug | 00:51 |
| jogo | clarkb: how do you think it should be fixed? | 00:51 |
| clarkb | I think pinning is bad | 00:51 |
| clarkb | capping is ok | 00:51 |
| mordred | greghaynes: https://github.com/emonty/deb-vhd-util <-- should be a git-buildpackage version of the thing that I pasted in earlier | 00:51 |
| * mordred disagrees with clarkb | 00:52 | |
| clarkb | jogo you want to pull in bugfixes for releases | 00:52 |
| clarkb | mordred yes I know | 00:52 |
| clarkb | so does sdague and jogo | 00:52 |
| greghaynes | mordred: what is this -s 0 -t 1, -s 1 -t 2 args to the same command? Is there just some jankyness to vhd-util where it requires two passes? | 00:52 |
| clarkb | which is fine yall are doing the work heee | 00:52 |
| mordred | I think all bugfix pulls should be explicit - because there are literally no python devs who don't break everything all the time | 00:52 |
| mordred | clarkb: although I _WANT_ to agree with you | 00:52 |
| mordred | because you're right | 00:52 |
| jogo | clarkb: so actually had a good chat with dhellmann about this | 00:52 |
| mordred | greghaynes: yes. | 00:52 |
| mordred | greghaynes: I have quite literally no idea what that does | 00:53 |
| clarkb | mordred so the problem with that attitude is breaking is a stick to get the fix | 00:53 |
| greghaynes | awesome | 00:53 |
| clarkb | no pins will ever be bumped in your prescribed world | 00:53 |
| jogo | and we will use ~= (actually a cap, allows for compatible changes) for oslo libs | 00:53 |
| jogo | for kilo | 00:53 |
| mordred | clarkb: fine by me | 00:53 |
| clarkb | jogo so thats a cap not a pin... | 00:53 |
| clarkb | jogo very different | 00:53 |
| jogo | clarkb: right, that is what I said. | 00:54 |
| jogo | wait I am confused | 00:54 |
| * mordred also agrees with clarkb's statement that mordred is not doing the work - so will honestly just occasionally say things but agree with anyone stepping up to the plate | 00:54 | |
| clarkb | jogo earlier you said pin everything | 00:54 |
| jogo | except for oslo libs | 00:54 |
| clarkb | jogo I said dont do that cap instead | 00:54 |
| clarkb | ~= is a cap | 00:54 |
| mordred | greghaynes: oh! btw - it does an in-place transform which we should probably figure out how to kill | 00:54 |
| mordred | greghaynes: as it kills the raw file | 00:54 |
| clarkb | oslo is no better or worse than * and should be treated the same | 00:55 |
| greghaynes | bahaha | 00:55 |
| *** vigneshvar has quit IRC | 00:55 | |
| mordred | greghaynes: OR - we should copy the raw file first | 00:55 |
| jogo | clarkb: use <= currently installed version, for everything but oslo libs. where we use '~='. Since we trust oslo to not break things for is | 00:55 |
| jogo | and when they do we move to | 00:55 |
| greghaynes | yes, ill just copy it for now | 00:55 |
| jogo | '<=' | 00:55 |
| jogo | anyway | 00:55 |
| jogo | yeah | 00:55 |
| clarkb | jogo I dont see reason for the distinction | 00:55 |
| clarkb | a dep is a dep | 00:56 |
| clarkb | I dont trust oslo | 00:56 |
| jogo | clarkb: oslo lib owners know about are release cycle and will maintain stable branches etc. | 00:56 |
| clarkb | history says oslo will break stable just as ofteb as everything else | 00:56 |
| jogo | that is completely different though | 00:56 |
| clarkb | why? | 00:56 |
| jogo | anyway yeah, unless you will do the work ... | 00:57 |
| *** yamamoto has quit IRC | 00:57 | |
| clarkb | its a single global list of deps | 00:57 |
| clarkb | any one can break you | 00:57 |
| jogo | clarkb: those weren't compatible dot releases | 00:57 |
| clarkb | oslo by definition isnt special here | 00:57 |
| jogo | (0.1.0 vs 0.2.0) | 00:57 |
| jogo | (0.1.0 vs 0.2.0 or 0.1.0 vs 0.1.1) | 00:57 |
| jogo | so take that up with dhellmann | 00:58 |
| jogo | but lets revisit this in a bit | 00:58 |
| jogo | want to hear more from tchaypo | 00:58 |
| clarkb | jogo no I think go the other routr | 00:58 |
| jogo | as I am hoping to sucker him into doing work :) | 00:58 |
| clarkb | use ~= on all deps | 00:58 |
| jogo | clarkb: so I have a hunch others share your view. so why not respond to the ML thread sigmavirus24 started | 00:59 |
| clarkb | jogo because its a day off :) | 00:59 |
| mordred | greghaynes: btw - this whole thing is so bonghits | 00:59 |
| clarkb | I havent read the ML today | 00:59 |
| clarkb | mordred YES | 01:00 |
| tchaypo | tchaypo is confused by this discussion | 01:00 |
| jogo | clarkb: I know. why is everyone here | 01:00 |
| tchaypo | tchaypo has nothing to say right now | 01:00 |
| jogo | tchaypo: just ignore what we said | 01:00 |
| *** Longgeek has quit IRC | 01:00 | |
| greghaynes | mordred: oh, I can tell. I was just thinking of how I miss the days when getting software to work required hunting down a random patch on some mailing list | 01:01 |
| jogo | tchaypo: TL;DR do we allow room for compatible releases in the requirements file | 01:01 |
| greghaynes | In an I totally wasnt thinking that kind of way | 01:01 |
| tchaypo | oh. I think the answer to that is no | 01:02 |
| jogo | tchaypo: so say pip-freeze says foo==1.0.0, in global requirements do we say: foo<=1.0.0 or foo<=1.1.0 | 01:02 |
| jogo | tchaypo: so say pip-freeze says foo==1.0.0, in global requirements do we say: foo<=1.0.0 or foo<1.1.0* | 01:02 |
| tchaypo | for master, sure. for stable, the requirements should be “this exact config is known to be stable" | 01:02 |
| mordred | greghaynes: :) | 01:02 |
| tchaypo | we can’t possibly assert that versions that aren’t released yet will be stable | 01:03 |
| jogo | tchaypo: right, so the risk there is if a bug fix comes out stable-maint must semi manually accept the new release | 01:03 |
| mordred | yes | 01:03 |
| tchaypo | yes | 01:03 |
| tchaypo | that makes sense to me. | 01:03 |
| mordred | because we have seen ZERO examples of the positive thing happening | 01:03 |
| mordred | and 10000000 examples of the negative thing happening | 01:03 |
| mordred | the "what if a bug fix comes out" is a complete myth | 01:03 |
| mordred | it has literally never happened and not broken something | 01:03 |
| jogo | mordred: who fixes bugs, we just call them features | 01:03 |
| mordred | jogo: ++ | 01:04 |
| *** ddieterl_ has quit IRC | 01:04 | |
| mordred | I mean, the only think worse than python releases are jenkins releases | 01:04 |
| tchaypo | If pip-freeze says foo==1.0.0, shouldn’t global-requirements say foo==1.0.0? | 01:04 |
| mordred | tchaypo: there is a camp of us who believe that | 01:04 |
| clarkb | mordred pbr did it recently | 01:04 |
| tchaypo | I wouldn’t want to say <=1.0.0 because I don’t know what features are missing from 0.9.9 | 01:04 |
| clarkb | mordred so the positive case happens | 01:04 |
| tchaypo | and I wouldn’t want to say <1.1.0 because I don’t know what bugs 1.0.1 will have | 01:04 |
| *** sdake has joined #openstack-infra | 01:05 | |
| mordred | clarkb: sure. but maybe if the stable-maint team didn't have to spend 100% of their time chasing the negative cases, they could positively bump requirements from time to time | 01:05 |
| clarkb | possibly | 01:05 |
| mordred | also - if nobody in stable got that pbr release, neat | 01:05 |
| mordred | unless someone needed it, in which case they could propose it | 01:05 |
| *** markvoelker has joined #openstack-infra | 01:05 | |
| jogo | tchaypo: so global reqs is used for two semi distinct things | 01:06 |
| mordred | this: | 01:06 |
| mordred | rmdir: failed to remove ‘debian/vhd-util/usr/local/share/doc/xen’: Directory not empty | 01:06 |
| mordred | is why debian packaging sucks | 01:06 |
| mordred | I DON'T CARE | 01:06 |
| jogo | tchaypo: 1) specify versions we test in the gate. 2) signal to downstream consumers what versions we are compatible with | 01:06 |
| mordred | I JUST WANT YOU TO TAKE THE THING I JUST RAN CONFIGURE MAKE MAKE INSTALL ON AND INSTALL IT AND DO NOTHING ELSE!!!!!! | 01:06 |
| jogo | tchaypo: from the gate point of view I think using '==' makes the most sense | 01:06 |
| mordred | making me figure out why usr/local/share/doc/xen has a file in it is a complete and total waste of time | 01:07 |
| jogo | tchaypo: from the downstream consumer point of view, I may want to test and validate a different set of deps. But I want to know what range upstream *thinks* maybe safe | 01:07 |
| tchaypo | my understanding is that == isn’t a complete fix (because of possible conflicts with system packages) but it’s a neccessary part of the fix | 01:07 |
| jogo | tchaypo: but really that distinction of which to use should be made by stable-maint team | 01:07 |
| clarkb | tchaypo and turing complete packaging | 01:07 |
| clarkb | but it does help quite a bit | 01:07 |
| tchaypo | yeah, downstreams should feel free to maintain their own patch on top of the requirements if they want :) | 01:07 |
| *** sdake_ has joined #openstack-infra | 01:08 | |
| tchaypo | clarkb: well yes. that too. | 01:08 |
| clarkb | tchaypo no push them up | 01:08 |
| tchaypo | has anyone implemented a rubix cube solver in setup.py yet? | 01:08 |
| jogo | tchaypo: so see my commit message https://github.com/openstack/requirements/commit/499db6b64071c2afa16390ad2bc024e6a96db4ff | 01:08 |
| jogo | tchaypo: and this https://github.com/openstack/requirements/commit/499db6b64071c2afa16390ad2bc024e6a96db4ff#diff-19163fe3a94f2bd3c8227ad62f5d069aR23 | 01:08 |
| jogo | this is the insanity that is deps | 01:09 |
| clarkb | my favorite setup.py thing recently was someone showed me package that opened a link in your browser upon install | 01:09 |
| jogo | pip-freeze says suds 0.4.1 | 01:09 |
| jogo | which isn't even out yet | 01:09 |
| jogo | but ubuntu has it | 01:09 |
| jogo | clarkb: amazing | 01:09 |
| *** weshay has quit IRC | 01:10 | |
| *** woodster_ has quit IRC | 01:10 | |
| tchaypo | "Grenade and tempest-full pip-freezes are different and do not work with each other" | 01:10 |
| tchaypo | lolwut | 01:10 |
| jogo | yup! | 01:10 |
| jogo | pretty fun isn't it | 01:10 |
| greghaynes | mordred: so, does it really matter that vhd-util modified the raw in place? AIUI the only reason we re-convert images even when we dont need to is to save space | 01:10 |
| clarkb | https://github.com/willkg/rdalal/blob/master/setup.py | 01:10 |
| *** markvoelker has quit IRC | 01:11 | |
| greghaynes | the caching shouldnt be an issue because we cache the original rootfs tarball/cloudimage, not the converted raw (I think) | 01:11 |
| jogo | and we use a lower ibvirt-python then we allow in g-r | 01:11 |
| tchaypo | do you mean that “pip freeze in these environments results in different output from the same input”? | 01:11 |
| *** sdake has quit IRC | 01:11 | |
| jogo | tchaypo: correct. well very similar inputs | 01:11 |
| jogo | both stable/juno packages | 01:12 |
| tchaypo | or do you meant that we need different dependencies because of other dependencies (eg, the issue where tempest itself was requiring a different version of something than the tools it was testing required)? | 01:12 |
| jogo | err openstack synced with g-r | 01:12 |
| jogo | tchaypo: tempest is now self contained | 01:12 |
| jogo | so it can require anything and won't break the world | 01:12 |
| tchaypo | re suds 0.4.1 - do we know how that happened? was it an issue where 0.4.1 release got pulled but the distros had already distributed it, or was it something else? | 01:12 |
| jogo | tchaypo: download http://logs.openstack.org/40/154740/1/check/check-tempest-dsvm-neutron-full/c232cfe/logs/pip-freeze.txt.gz | 01:13 |
| jogo | and the grenade version of that pip freeze | 01:13 |
| tchaypo | clarkb: that’s nuts. | 01:13 |
| jogo | and diff them | 01:13 |
| jogo | tchaypo: for suds all I know is this: | 01:13 |
| jogo | https://pypi.python.org/pypi/suds | 01:13 |
| jogo | http://packages.ubuntu.com/search?keywords=python-suds | 01:14 |
| adam_g | wtf? | 01:14 |
| jogo | 0.4 http://changelogs.ubuntu.com/changelogs/pool/main/s/suds/suds_0.4.1-11build1/changelog | 01:15 |
| jogo | * Merging upstream version 0.4. | 01:15 |
| jogo | so they forked? | 01:15 |
| adam_g | debian merged a beta | 01:15 |
| adam_g | it looks like? | 01:15 |
| adam_g | * Merging upstream version 0.4.1. | 01:15 |
| jogo | adam_g: lol WTF | 01:16 |
| adam_g | -- Daniel Baumann <daniel@debian.org> Sun, 31 Oct 2010 16:10:08 +0100 | 01:16 |
| adam_g | https://fedorahosted.org/suds/ lists 0.4.1 Beta (NOT AVAILABLE) | 01:16 |
| jogo | adam_g: so this is another issue -- we use a terrible mix of pip and apt-get installed python packages | 01:16 |
| jogo | we need to stop that | 01:16 |
| *** jamielennox is now known as jamielennox|away | 01:16 | |
| mordred | yes | 01:16 |
| mordred | please god yes | 01:16 |
| clarkb | +1 to that | 01:17 |
| jogo | actually this shouldn't be too nard | 01:17 |
| jogo | hard | 01:17 |
| jogo | just delete lines from devstack | 01:17 |
| mordred | jogo: I've made that patch like 4 times | 01:17 |
| *** sigmavirus24 is now known as sigmavirus24_awa | 01:17 | |
| jogo | mordred: and? | 01:17 |
| clarkb | jogo mordred you have to install "optional" deps too | 01:17 |
| mordred | well, I either get shot down, or ignored, or somethign doesn't work | 01:17 |
| clarkb | so mysql-python, libvirt-python and so on | 01:18 |
| marun | +1 to pip only | 01:19 |
| adam_g | http://paste.ubuntu.com/10264990/ | 01:19 |
| marun | btw, I've started prototyping something that would figure out the os dependency graph of neutron's python dependencies | 01:20 |
| marun | it's a manual process, but uses docker to try to install each dep in a bare system with only specified distro packages | 01:20 |
| marun | it's ubuntu only for now but could be extended to be multi-distro easily enough | 01:20 |
| tchaypo | jogo: you also mentioned "more transparent reporting of stable branch statuses somehow. So its easier to see if its working not working etc.” - could you elaborate on that? | 01:20 |
| tchaypo | marun: I think that if I said “can’t you just use pip-compile for that?” and pointed you at http://nvie.com/posts/better-package-management/ | 01:21 |
| tchaypo | you’d respond “no, because that doesn’t help you capture the dependencies you aren’t noticing because the system has them already" | 01:21 |
| marun | manual meaning that I've had to deduce which os packages are required when failure occurs. but validating that a package installs with only specified deps is easy with docker | 01:21 |
| jogo | marun: neat! | 01:21 |
| tchaypo | does “os” in this instance mean “operating system” or “openstack”? | 01:21 |
| marun | tchaypo: ah, sorry | 01:22 |
| marun | distro pckages | 01:22 |
| *** signed8bit is now known as signed8bit_ZZZzz | 01:22 | |
| jogo | tchaypo: yeah, so make it easier for everyone to know what the current status of stable branches are | 01:22 |
| jogo | are they working, broken etc. | 01:22 |
| tchaypo | so a dashboard kind of thing? | 01:22 |
| adam_g | oh | 01:22 |
| jogo | that has been an issue when they break the world. stable maint people don't know right away | 01:22 |
| adam_g | speaking of which | 01:22 |
| adam_g | nova stable/juno is currently broken | 01:22 |
| jogo | tchaypo: yeah, not sure what it would look like. how that fits in other things | 01:22 |
| jogo | adam_g: ohh see I didn't even know that | 01:22 |
| adam_g | jogo, :) | 01:22 |
| adam_g | https://review.openstack.org/#/c/156410/ is needed there | 01:23 |
| marun | clarkb: does pip-compile install distro deps? | 01:23 |
| clarkb | no | 01:23 |
| clarkb | it just lists deps from pip | 01:23 |
| tchaypo | jogo: how do stable-maint people find out about the breakage? | 01:23 |
| marun | clarkb: so it's python-only deps? | 01:23 |
| clarkb | marun aiui yes | 01:23 |
| jogo | anyway heading out for a bit | 01:23 |
| tchaypo | I’m hoping the answer is “the daily job fails and emails” rather than “someone complains eventually" | 01:23 |
| marun | clarkb: ah, ok. That would make it easier to build the graph. | 01:24 |
| adam_g | tchaypo, both of those | 01:24 |
| jogo | bbiab | 01:24 |
| adam_g | tchaypo, the daily jobs only run the project's py2*, doc jobs for bitrot | 01:24 |
| marun | clarkb: there is still a need to figure out which distro deps are required by a given python dep (if any) | 01:24 |
| marun | clarkb: that's the main thing I was doing, the hard way as it turns out. pip-compile will make the process easier | 01:24 |
| tchaypo | marun: is your work somewhere I could take a look at it and see if I can contribute? | 01:24 |
| adam_g | tchaypo, when things like requirements break project wide tests like the tempest or grenade jobs, that usually doesnt get noticed until someone cant land code | 01:25 |
| marun | clarkb: I figure once it's done, we could add the relationship between python dep and distro dep to the requirements repo or something like that | 01:25 |
| clarkb | marun ya the mapping is the hard bit | 01:25 |
| marun | tchaypo: hah, no. it's a shell script, yaml file, and some docker images. | 01:26 |
| marun | tchaypo: so, very little. but if you want I could paste it up | 01:26 |
| *** baoli has quit IRC | 01:27 | |
| *** bhuvan has quit IRC | 01:27 | |
| *** sarob has joined #openstack-infra | 01:28 | |
| *** Longgeek has joined #openstack-infra | 01:28 | |
| *** Longgeek has quit IRC | 01:28 | |
| *** Longgeek has joined #openstack-infra | 01:30 | |
| *** ddieterly has joined #openstack-infra | 01:30 | |
| *** zhiwei has joined #openstack-infra | 01:30 | |
| *** jamielennox|away is now known as jamielennox | 01:31 | |
| tchaypo | marun: well, I’m a bit jetlagged right now so I won’t be doing anything with it tonight anyway | 01:31 |
| marun | tchaypo: I'll work on getting it up on github later this week then. | 01:32 |
| pleia2 | anteaya: thanks, honored to be considered among that list | 01:32 |
| tchaypo | and for tomorrow my goal is just to play with devstack | 01:32 |
| marun | tchaypo: feel free to poke me if I'm too slow. | 01:32 |
| tchaypo | that’d be cool :) | 01:32 |
| *** sarob has quit IRC | 01:32 | |
| *** tiswanso has quit IRC | 01:34 | |
| *** hdd has quit IRC | 01:38 | |
| *** yamamoto_ has joined #openstack-infra | 01:40 | |
| *** jamielennox is now known as jamielennox|away | 01:41 | |
| *** gulic has quit IRC | 01:42 | |
| *** bdpayne has quit IRC | 01:49 | |
| *** ddieterl_ has joined #openstack-infra | 01:51 | |
| *** ddieterly has quit IRC | 01:53 | |
| *** ddieterly has joined #openstack-infra | 01:54 | |
| *** ddieterl_ has quit IRC | 01:54 | |
| *** avozza is now known as zz_avozza | 01:56 | |
| *** gulic has joined #openstack-infra | 02:00 | |
| *** tobiash has quit IRC | 02:00 | |
| *** tobiash has joined #openstack-infra | 02:01 | |
| *** woodster_ has joined #openstack-infra | 02:03 | |
| *** gulic has quit IRC | 02:07 | |
| openstackgerrit | greghaynes proposed openstack-infra/storyboard: Use Python 3 compatible configparser https://review.openstack.org/155922 | 02:07 |
| *** harlowja_at_home has joined #openstack-infra | 02:07 | |
| *** markvoelker has joined #openstack-infra | 02:08 | |
| *** jamielennox|away is now known as jamielennox | 02:09 | |
| *** markvoelker has quit IRC | 02:13 | |
| *** tiswanso has joined #openstack-infra | 02:13 | |
| *** salv-orlando has quit IRC | 02:13 | |
| *** ddieterl_ has joined #openstack-infra | 02:18 | |
| openstackgerrit | greghaynes proposed openstack-infra/storyboard: Use TempDir fixture for working dir https://review.openstack.org/155956 | 02:20 |
| *** ddieterly has quit IRC | 02:21 | |
| openstackgerrit | Imran Hayder Malik proposed openstack-infra/jenkins-job-builder: Add details to credentials param in scm docs https://review.openstack.org/156141 | 02:24 |
| *** dmsimard_away is now known as dmsimard | 02:25 | |
| *** nelsnels_ has quit IRC | 02:26 | |
| *** Ryan_Lane has joined #openstack-infra | 02:28 | |
| *** sarob has joined #openstack-infra | 02:29 | |
| *** harlowja_at_home has quit IRC | 02:33 | |
| *** sarob has quit IRC | 02:34 | |
| openstackgerrit | greghaynes proposed openstack-infra/storyboard: Use TempDir fixture for working dir https://review.openstack.org/155956 | 02:35 |
| *** bknudson has quit IRC | 02:37 | |
| *** dims__ has quit IRC | 02:40 | |
| *** harlowja_at_home has joined #openstack-infra | 02:40 | |
| *** ddieterl_ has quit IRC | 02:41 | |
| *** ddieterly has joined #openstack-infra | 02:41 | |
| *** nelsnelson has joined #openstack-infra | 02:48 | |
| *** ddieterl_ has joined #openstack-infra | 02:52 | |
| *** ddieterly has quit IRC | 02:53 | |
| *** kaisers has joined #openstack-infra | 02:53 | |
| *** hichihara has quit IRC | 02:54 | |
| *** kaisers1 has quit IRC | 02:55 | |
| *** ddieterl_ has quit IRC | 02:56 | |
| *** ddieterly has joined #openstack-infra | 02:56 | |
| *** ghostpl_ has quit IRC | 02:58 | |
| *** tqtran has quit IRC | 03:02 | |
| *** zz_dimtruck is now known as dimtruck | 03:03 | |
| *** bhunter71 has quit IRC | 03:04 | |
| *** armax has joined #openstack-infra | 03:06 | |
| *** marun has quit IRC | 03:07 | |
| *** markvoelker has joined #openstack-infra | 03:09 | |
| *** sarob has joined #openstack-infra | 03:12 | |
| *** salv-orlando has joined #openstack-infra | 03:14 | |
| *** markvoelker has quit IRC | 03:14 | |
| *** Ryan_Lane has quit IRC | 03:15 | |
| *** nelsnelson has quit IRC | 03:15 | |
| *** sdake has joined #openstack-infra | 03:23 | |
| *** sdake_ has quit IRC | 03:26 | |
| openstackgerrit | Matt Riedemann proposed openstack-infra/elastic-recheck: Add query for ironic grenade fail 1422556 https://review.openstack.org/156453 | 03:32 |
| *** dims__ has joined #openstack-infra | 03:40 | |
| *** amotoki_ has joined #openstack-infra | 03:43 | |
| *** amotoki_ has quit IRC | 03:45 | |
| *** dims__ has quit IRC | 03:45 | |
| *** harlowja_at_home has quit IRC | 03:46 | |
| openstackgerrit | Steve Martinelli proposed openstack-infra/project-config: Run openstackclient functional tests against keystoneclient master https://review.openstack.org/156456 | 03:47 |
| *** sdake_ has joined #openstack-infra | 03:47 | |
| *** emagana has joined #openstack-infra | 03:51 | |
| *** sdake has quit IRC | 03:52 | |
| *** reed has quit IRC | 03:56 | |
| *** claudiub has quit IRC | 03:56 | |
| *** achanda has joined #openstack-infra | 03:56 | |
| *** eharney has quit IRC | 03:56 | |
| *** koolhead17 has joined #openstack-infra | 03:59 | |
| *** Longgeek has quit IRC | 04:04 | |
| *** eharney has joined #openstack-infra | 04:10 | |
| *** markvoelker has joined #openstack-infra | 04:10 | |
| *** ddieterly has quit IRC | 04:11 | |
| *** BharatK has joined #openstack-infra | 04:13 | |
| *** otter768 has quit IRC | 04:13 | |
| *** markvoelker has quit IRC | 04:16 | |
| openstackgerrit | Ian Wienand proposed openstack-infra/system-config: Use nodepool 'generate-log-config' command to generate log config https://review.openstack.org/156457 | 04:17 |
| openstackgerrit | Ian Wienand proposed openstack-infra/nodepool: Add generate-log-config command https://review.openstack.org/153904 | 04:18 |
| openstackgerrit | greghaynes proposed openstack-infra/storyboard: Fix bytes vs string python 3 issues https://review.openstack.org/155969 | 04:22 |
| *** sdake has joined #openstack-infra | 04:26 | |
| *** ghostpl_ has joined #openstack-infra | 04:29 | |
| *** sdake_ has quit IRC | 04:30 | |
| *** coolsvap_ is now known as coolsvap | 04:31 | |
| *** yamamoto_ has quit IRC | 04:33 | |
| *** sdake_ has joined #openstack-infra | 04:33 | |
| *** ghostpl_ has quit IRC | 04:35 | |
| *** yamamoto_ has joined #openstack-infra | 04:35 | |
| openstackgerrit | greghaynes proposed openstack-infra/storyboard: Fix bytes vs string python 3 issues https://review.openstack.org/155969 | 04:36 |
| *** jerryz has joined #openstack-infra | 04:37 | |
| *** emagana_ has joined #openstack-infra | 04:37 | |
| *** sdake has quit IRC | 04:37 | |
| *** tiswanso has quit IRC | 04:38 | |
| *** spzala has quit IRC | 04:40 | |
| *** emagana has quit IRC | 04:40 | |
| *** koolhead17 has quit IRC | 04:42 | |
| *** koolhead17 has joined #openstack-infra | 04:43 | |
| *** deva__ has joined #openstack-infra | 04:47 | |
| *** hdd has joined #openstack-infra | 04:50 | |
| *** koolhead_ has joined #openstack-infra | 04:52 | |
| *** deva__ has quit IRC | 04:54 | |
| *** devanand_ has joined #openstack-infra | 04:54 | |
| *** koolhead17 has quit IRC | 04:55 | |
| *** sdake has joined #openstack-infra | 04:57 | |
| *** sdake_ has quit IRC | 05:00 | |
| *** radez is now known as radez_g0n3 | 05:01 | |
| *** coolsvap is now known as coolsvap_ | 05:02 | |
| *** dmsimard is now known as dmsimard_away | 05:03 | |
| *** sdake has quit IRC | 05:08 | |
| *** coolsvap_ is now known as coolsvap | 05:09 | |
| *** sdake has joined #openstack-infra | 05:10 | |
| *** nikil89_ has joined #openstack-infra | 05:11 | |
| *** nikil89_ is now known as naggappan | 05:11 | |
| *** ddieterly has joined #openstack-infra | 05:12 | |
| *** markvoelker has joined #openstack-infra | 05:12 | |
| *** andrey-mp has joined #openstack-infra | 05:14 | |
| andrey-mp | please, can someone review this - https://review.openstack.org/#/c/156293/ ? | 05:14 |
| *** yamahata has joined #openstack-infra | 05:14 | |
| *** ilyashakhat has joined #openstack-infra | 05:16 | |
| *** devlaps has quit IRC | 05:16 | |
| *** ddieterly has quit IRC | 05:16 | |
| *** markvoelker has quit IRC | 05:17 | |
| *** koolhead17 has joined #openstack-infra | 05:18 | |
| *** sdake_ has joined #openstack-infra | 05:20 | |
| *** koolhead_ has quit IRC | 05:20 | |
| *** sdake__ has joined #openstack-infra | 05:22 | |
| *** sdake has quit IRC | 05:22 | |
| openstackgerrit | Steven Dake proposed openstack-infra/project-config: Add docs to the magnum gate https://review.openstack.org/149334 | 05:24 |
| openstackgerrit | Steven Dake proposed openstack-infra/project-config: Add docs to the magnum gate https://review.openstack.org/149334 | 05:24 |
| *** sdake_ has quit IRC | 05:25 | |
| *** andrey-mp has quit IRC | 05:27 | |
| *** dimtruck is now known as zz_dimtruck | 05:29 | |
| *** yfried has quit IRC | 05:31 | |
| *** sdake has joined #openstack-infra | 05:32 | |
| *** ilyashakhat has quit IRC | 05:35 | |
| *** sdake__ has quit IRC | 05:36 | |
| openstackgerrit | Merged openstack-infra/project-config: Use correct regex for EC2 tempest gating https://review.openstack.org/156293 | 05:36 |
| *** koolhead_ has joined #openstack-infra | 05:40 | |
| *** achanda has quit IRC | 05:43 | |
| *** koolhead17 has quit IRC | 05:44 | |
| *** hdd has quit IRC | 05:44 | |
| *** achanda has joined #openstack-infra | 05:46 | |
| *** ilyashakhat has joined #openstack-infra | 05:46 | |
| *** pcaruana has quit IRC | 05:47 | |
| openstackgerrit | Khai Do proposed openstack-infra/project-config: Make gerritbot report changes from Gerrit 2.9.4 branch https://review.openstack.org/155450 | 05:48 |
| *** sdake_ has joined #openstack-infra | 05:52 | |
| *** sarob has quit IRC | 05:52 | |
| *** davideagnello has quit IRC | 05:53 | |
| *** coolsvap is now known as coolsvap_ | 05:53 | |
| *** sdake__ has joined #openstack-infra | 05:54 | |
| *** sdake has quit IRC | 05:54 | |
| *** sdake_ has quit IRC | 05:58 | |
| *** r1chardj0n3s is now known as r1chardj0n3s_afk | 05:58 | |
| *** coolsvap_ is now known as coolsvap | 06:00 | |
| *** jcoufal has joined #openstack-infra | 06:05 | |
| *** zhiwei has left #openstack-infra | 06:07 | |
| *** dmsimard_away has quit IRC | 06:09 | |
| *** zhiwei has joined #openstack-infra | 06:09 | |
| *** dmsimard_away has joined #openstack-infra | 06:09 | |
| *** dmsimard_away is now known as dmsimard | 06:09 | |
| *** zhiwei has left #openstack-infra | 06:10 | |
| *** ddieterly has joined #openstack-infra | 06:13 | |
| *** markvoelker has joined #openstack-infra | 06:13 | |
| *** otter768 has joined #openstack-infra | 06:14 | |
| *** cipcosma has joined #openstack-infra | 06:17 | |
| *** ddieterly has quit IRC | 06:17 | |
| *** otter768 has quit IRC | 06:19 | |
| *** ghostpl_ has joined #openstack-infra | 06:19 | |
| *** markvoelker has quit IRC | 06:19 | |
| *** AJaeger has joined #openstack-infra | 06:20 | |
| *** ghostpl_ has quit IRC | 06:25 | |
| *** devanand_ has quit IRC | 06:28 | |
| *** zz_avozza is now known as avozza | 06:31 | |
| *** xyang1 has quit IRC | 06:33 | |
| *** vigneshvar has joined #openstack-infra | 06:34 | |
| *** ilyashakhat has quit IRC | 06:43 | |
| *** sdake__ has quit IRC | 06:54 | |
| *** yamamoto_ has quit IRC | 06:57 | |
| *** sarob has joined #openstack-infra | 06:57 | |
| *** kaisers has quit IRC | 06:57 | |
| *** jerryz has quit IRC | 06:59 | |
| *** yamahata has quit IRC | 07:00 | |
| *** sarob has quit IRC | 07:04 | |
| *** redixin has quit IRC | 07:06 | |
| *** salv-orlando has quit IRC | 07:07 | |
| *** yfried has joined #openstack-infra | 07:09 | |
| *** yamamoto_ has joined #openstack-infra | 07:10 | |
| *** yamamoto_ has quit IRC | 07:12 | |
| *** yfried_ has joined #openstack-infra | 07:13 | |
| *** ddieterly has joined #openstack-infra | 07:13 | |
| *** yfried has quit IRC | 07:14 | |
| *** markvoelker has joined #openstack-infra | 07:15 | |
| *** camunoz is now known as camunoz_gone | 07:16 | |
| *** ddieterly has quit IRC | 07:18 | |
| *** dims__ has joined #openstack-infra | 07:18 | |
| *** yfried_ is now known as yfried|afk | 07:19 | |
| *** unicell has joined #openstack-infra | 07:19 | |
| *** scheuran has joined #openstack-infra | 07:19 | |
| *** funzo_ has joined #openstack-infra | 07:20 | |
| *** markvoelker has quit IRC | 07:21 | |
| *** funzo has quit IRC | 07:22 | |
| *** dims__ has quit IRC | 07:23 | |
| *** mrunge has joined #openstack-infra | 07:23 | |
| *** yfried|afk is now known as yfried_ | 07:25 | |
| *** sputnik13 has joined #openstack-infra | 07:30 | |
| *** ZZelle has quit IRC | 07:32 | |
| *** ZZelle has joined #openstack-infra | 07:32 | |
| *** yfried_ is now known as yfried|afk | 07:35 | |
| *** stevemar has quit IRC | 07:38 | |
| *** Ala has joined #openstack-infra | 07:40 | |
| *** jgallard__ has joined #openstack-infra | 07:40 | |
| *** ilyashakhat has joined #openstack-infra | 07:44 | |
| *** ghostpl_ has joined #openstack-infra | 07:45 | |
| *** mrmartin has joined #openstack-infra | 07:45 | |
| *** achanda has quit IRC | 07:47 | |
| openstackgerrit | yolanda.robla proposed openstack-infra/infra-specs: Spec for nodepool image update retries https://review.openstack.org/155305 | 07:47 |
| *** ilyashakhat has quit IRC | 07:48 | |
| *** mrmartin has quit IRC | 07:49 | |
| *** mrmartin has joined #openstack-infra | 07:50 | |
| *** ghostpl_ has quit IRC | 07:50 | |
| *** yfried|afk is now known as yfried_ | 07:53 | |
| *** mkerrin has joined #openstack-infra | 07:54 | |
| *** achanda has joined #openstack-infra | 07:54 | |
| *** devvesa has joined #openstack-infra | 07:54 | |
| *** Ala has quit IRC | 07:56 | |
| *** rm_work is now known as rm_work|away | 07:57 | |
| *** Ala has joined #openstack-infra | 08:01 | |
| *** HeOS has quit IRC | 08:03 | |
| *** masoz has joined #openstack-infra | 08:03 | |
| devvesa | good morning. could a core dev do a quick review on this patch? https://review.openstack.org/#/c/156268/ | 08:04 |
| devvesa | i would appreciate a lot, cause it blocks our development | 08:04 |
| *** andreykurilin_ has joined #openstack-infra | 08:05 | |
| *** fifieldt has joined #openstack-infra | 08:09 | |
| *** belmoreira has joined #openstack-infra | 08:09 | |
| *** chlong has quit IRC | 08:11 | |
| *** unicell has quit IRC | 08:13 | |
| *** belmoreira has quit IRC | 08:14 | |
| *** ddieterly has joined #openstack-infra | 08:14 | |
| *** otter768 has joined #openstack-infra | 08:15 | |
| *** markus_z has joined #openstack-infra | 08:15 | |
| amotoki | hi infra team, could you add me (amotoki@gerrit) to the inital gerrit groups for our stackforge projets? the group names are networking-nec-* | 08:16 |
| *** markvoelker has joined #openstack-infra | 08:17 | |
| *** arxcruz has joined #openstack-infra | 08:19 | |
| *** ddieterly has quit IRC | 08:19 | |
| *** cody-somerville has quit IRC | 08:19 | |
| *** otter768 has quit IRC | 08:20 | |
| *** achanda has quit IRC | 08:20 | |
| *** markvoelker has quit IRC | 08:22 | |
| *** dtantsur|afk is now known as dtantsur | 08:23 | |
| *** achanda has joined #openstack-infra | 08:24 | |
| gilliard | Does anyone else find underscore characters to be invisible in gerrit's code view? | 08:25 |
| *** sarob has joined #openstack-infra | 08:25 | |
| AJaeger | devvesa: please ping again at US time. I already +2ed and I'm not sure whether jhesketh or SergeyLukjanov are around to review 156268 | 08:28 |
| openstackgerrit | yolanda.robla proposed openstack-infra/infra-specs: Add nodepool image testing spec. https://review.openstack.org/139598 | 08:28 |
| AJaeger | amotoki: I'm sure one of the admins will add amotiki@gerrit to networking-nic groups in Jenkins. Let's ask fungi or clarkb and hope for scroll back ;) | 08:29 |
| AJaeger | amotoki: both are US based, so will take a few hours. | 08:29 |
| AJaeger | gilliard: do you have an example review to look at? | 08:29 |
| *** achanda has quit IRC | 08:30 | |
| devvesa | AJaeger: ok. I'll try again this afternoon | 08:30 |
| devvesa | ('my' afternoon :) ) | 08:30 |
| *** Guest10052 has joined #openstack-infra | 08:30 | |
| *** sarob has quit IRC | 08:31 | |
| AJaeger | devvesa: If you're also European based, it's our afternoon ;) | 08:32 |
| *** achuprin_ has quit IRC | 08:33 | |
| anteaya | good afternoon | 08:37 |
| *** achanda has joined #openstack-infra | 08:37 | |
| anteaya | mikal: prepatory ping to you for nova-net to neutron migration meeting in 23 minutes | 08:37 |
| gilliard | AJaeger: I'm looking at https://review.openstack.org/#/c/151664/2/nova/virt/libvirt/driver.py,cm line 5346 | 08:38 |
| *** salv-orlando has joined #openstack-infra | 08:38 | |
| AJaeger | good morning, anteaya. What timezone are you in? | 08:38 |
| anteaya | AJaeger: eastern | 08:39 |
| *** ghostpl_ has joined #openstack-infra | 08:39 | |
| AJaeger | anteaya do you have time to help devvesa with approving https://review.openstack.org/#/c/156268/ ? | 08:39 |
| anteaya | the middle of the night meetings | 08:39 |
| AJaeger | anteaya: time to sleep! | 08:39 |
| AJaeger | ;( | 08:39 |
| anteaya | after meetings | 08:39 |
| gilliard | AJaeger: http://i.imgur.com/2e0WqfY.png | 08:39 |
| AJaeger | gilliard: this displays just fine in firefox for me | 08:39 |
| AJaeger | gilliard: in 5350 you have the underscores but not in 5346 | 08:40 |
| anteaya | gilliard: I'm in firefox I can not see the underscores on the line you mention | 08:40 |
| anteaya | oh | 08:40 |
| AJaeger | gilliard: try without colours: https://review.openstack.org/#/c/151664/2/nova/virt/libvirt/driver.py | 08:40 |
| anteaya | AJaeger: I confirm what you see | 08:41 |
| AJaeger | gilliard: works in chromium nicely as well for me | 08:41 |
| gilliard | so I see 'VIR DOMAIN JOB FAILED' and yous all see 'VIR_DOMAIN_JOB_FAILED' ? | 08:41 |
| AJaeger | gilliard: correct | 08:42 |
| anteaya | hang on | 08:42 |
| gilliard | OK no worries. I'll just change the font a bit with a script. It's usually pretty obvious where the _ are. | 08:42 |
| anteaya | in line 5346 of the view you posted gilliard I do not see the underscores | 08:42 |
| *** MaxV has joined #openstack-infra | 08:43 | |
| anteaya | gilliard: why are you using that view? | 08:43 |
| anteaya | can you use the old view that AJaeger posted? | 08:43 |
| AJaeger | yes, the screenshot does not show it | 08:43 |
| anteaya | gilliard: I believe you are using the new view in gerrit | 08:43 |
| anteaya | you can set it in settings | 08:43 |
| gilliard | yes I changed at the nova midcycle because it seems a bit nicer on the whole-patch view. but an individual file view has this problem. | 08:43 |
| anteaya | ah | 08:45 |
| gilliard | The old view is fine http://i.imgur.com/1SBrjMR.png | 08:45 |
| anteaya | gilliard: yes that is what I see | 08:46 |
| *** achanda has quit IRC | 08:47 | |
| anteaya | devvesa: so do you need to keep pypi jobs on your repo or not? Your commit message says you don't need them yet they are still in http://git.openstack.org/cgit/openstack-infra/project-config/tree/jenkins/jobs/projects.yaml#n3743 as well as the zuul/layout.yaml file | 08:47 |
| anteaya | devvesa: if you don't need pypi jobs can you remove them in both files please? | 08:47 |
| *** achuprin_ has joined #openstack-infra | 08:47 | |
| *** ivar-lazzaro has quit IRC | 08:47 | |
| openstackgerrit | Bob Ball proposed openstack-infra/nodepool: Add support for requsting that config-drive be used. https://review.openstack.org/155770 | 08:47 |
| gilliard | You can see from L#5350 that it's not happening on every _ char, so I figure it's a rendering bug in firefox, but no harm in asking :) | 08:47 |
| anteaya | I find it concerning that one of gerrit's views is changing the code | 08:48 |
| devvesa | anteaya: oh. that's a typo error. I meant pypy | 08:48 |
| anteaya | I consider removing underscores from the view to be changing the code | 08:48 |
| *** sdake has joined #openstack-infra | 08:48 | |
| anteaya | devvesa: can you update the commit message please? | 08:48 |
| devvesa | sure | 08:48 |
| gilliard | Yes I've almost -1'd a couple of patches because of what I thought must be typos. | 08:48 |
| anteaya | gilliard: that is concerning | 08:49 |
| openstackgerrit | Jaume Devesa proposed openstack-infra/project-config: Remove jobs for networking-midonet patches https://review.openstack.org/156268 | 08:49 |
| devvesa | anteaya: AJaeger: done | 08:49 |
| anteaya | gilliard: do you know if there is any bug filed to gerrit about it? | 08:49 |
| anteaya | not sure if it is a gerrit issue or browser issue | 08:49 |
| gilliard | I'll have a look. | 08:50 |
| anteaya | but if the old view doesnt' do it and the new view does, I would start with gerrit | 08:50 |
| anteaya | gilliard: thanks | 08:50 |
| amotoki | AJaeger: thanks. I know it is mainly US based. I wil request it later. | 08:50 |
| gilliard | http://code.google.com/p/gerrit/issues/detail?id=1734&q=underscore&colspec=ID%20Type%20Stars%20Milestone%20Status%20Priority%20Owner%20Summary | 08:50 |
| *** jistr has joined #openstack-infra | 08:51 | |
| devvesa | anteaya: thanks | 08:51 |
| gilliard | sorry, anteaya, better link: http://code.google.com/p/gerrit/issues/detail?id=1734 | 08:52 |
| gilliard | Bug is over 2 years old... | 08:52 |
| anteaya | awesome | 08:52 |
| anteaya | I don't know what to suggest | 08:52 |
| gilliard | I'll get over it :) | 08:53 |
| anteaya | does the description match your experience? | 08:53 |
| anteaya | gilliard: :) | 08:53 |
| *** jlibosva has joined #openstack-infra | 08:53 | |
| *** mkerrin has quit IRC | 08:53 | |
| gilliard | perhaps I'll start by removing all the underscores from nova. | 08:54 |
| anteaya | gilliard: hey, there's an idea | 08:55 |
| anteaya | sure to have a beneficial outcome | 08:55 |
| *** cody-somerville has joined #openstack-infra | 08:56 | |
| *** fifieldt has quit IRC | 08:56 | |
| AJaeger | devvesa: I'll approve your patch soon | 08:56 |
| AJaeger | amotoki: I hope one of the admins scrolls back the chat channel and just does the right thing ;) | 08:57 |
| *** jerryz has joined #openstack-infra | 08:57 | |
| devvesa | AJaeger: awesome. Thanks for taking care of this | 08:58 |
| AJaeger | devvesa: you're welcome! | 08:59 |
| *** jp_at_hp has joined #openstack-infra | 08:59 | |
| *** sdake_ has joined #openstack-infra | 09:00 | |
| openstackgerrit | Merged openstack-infra/project-config: Remove jobs for networking-midonet patches https://review.openstack.org/156268 | 09:03 |
| *** sdake has quit IRC | 09:03 | |
| *** jedimike has joined #openstack-infra | 09:05 | |
| AJaeger | devvesa: Now the setup needs to be updated, if you run recheck on your change in half an hour, the jobs should be removed. | 09:05 |
| devvesa | AJaeger: Gotcha! Thanks | 09:07 |
| devvesa | AJaeger: there is a job in the 'post' pipeline called 'publish-specs-site' in zuul | 09:08 |
| devvesa | do I have to wait this one? | 09:08 |
| AJaeger | no need to wait for that one | 09:09 |
| devvesa | ok. | 09:09 |
| AJaeger | these are independent | 09:09 |
| *** Ala has quit IRC | 09:10 | |
| *** Ala has joined #openstack-infra | 09:12 | |
| *** Ala has quit IRC | 09:12 | |
| *** Ala has joined #openstack-infra | 09:13 | |
| *** salv-orlando has quit IRC | 09:15 | |
| *** ddieterly has joined #openstack-infra | 09:15 | |
| *** mase_home has quit IRC | 09:16 | |
| anteaya | jhesketh: are you in a postion to rouse gus and mikal? | 09:16 |
| anteaya | jhesketh: they are awol at the migration meeting and I kind of need them both | 09:16 |
| *** markvoelker has joined #openstack-infra | 09:18 | |
| *** HeOS has joined #openstack-infra | 09:19 | |
| *** ddieterly has quit IRC | 09:19 | |
| *** spredzy|afk is now known as spredzy | 09:21 | |
| *** markvoelker has quit IRC | 09:23 | |
| *** nfedotov1 has joined #openstack-infra | 09:23 | |
| *** skolekonov has joined #openstack-infra | 09:25 | |
| *** nfedotov has quit IRC | 09:25 | |
| *** avozza is now known as zz_avozza | 09:26 | |
| *** ihrachyshka has joined #openstack-infra | 09:26 | |
| *** ilyashakhat has joined #openstack-infra | 09:27 | |
| *** sarob has joined #openstack-infra | 09:28 | |
| jhesketh | anteaya: outside of pinging them like you just did I hold no special powers sorry | 09:28 |
| jhesketh | I'll send them a text though | 09:28 |
| anteaya | jhesketh: thank you | 09:28 |
| anteaya | jhesketh: oh that may help | 09:28 |
| anteaya | danke | 09:28 |
| jhesketh | anteaya: unfortunately mikal apologises he will not be able to attend | 09:30 |
| anteaya | jhesketh: thanks for texting him and relaying the message | 09:31 |
| *** yfried_ is now known as yfried|afk | 09:31 | |
| anteaya | he did inform me it was right in the middle of putting children to bed | 09:31 |
| anteaya | but I was hopeful | 09:31 |
| *** ilyashakhat has quit IRC | 09:31 | |
| *** sarob has quit IRC | 09:36 | |
| jhesketh | yeah, he's also been up since incredibly early | 09:38 |
| *** yfried|afk is now known as yfried_ | 09:38 | |
| openstackgerrit | David Pursehouse proposed openstack-infra/jenkins-job-builder: Fix wrong casing in JDKAxis <name> tag https://review.openstack.org/156519 | 09:39 |
| anteaya | jhesketh: ah yeah that would do it too | 09:39 |
| openstackgerrit | Merged openstack-infra/storyboard-webclient: Fix error notifications https://review.openstack.org/155242 | 09:41 |
| *** ssam2 has joined #openstack-infra | 09:41 | |
| *** dizquierdo has joined #openstack-infra | 09:42 | |
| openstackgerrit | Joshua Hesketh proposed openstack-infra/project-config: Remove jenkins console-log publisher from infra https://review.openstack.org/156520 | 09:42 |
| openstackgerrit | Joshua Hesketh proposed openstack-infra/project-config: Enable swift logs for python jobs https://review.openstack.org/156521 | 09:42 |
| openstackgerrit | Joshua Hesketh proposed openstack-infra/project-config: Upload python test results to swift https://review.openstack.org/156522 | 09:42 |
| *** mpaolino has joined #openstack-infra | 09:45 | |
| *** mpaolino has quit IRC | 09:46 | |
| *** e0ne has joined #openstack-infra | 09:48 | |
| *** zz_avozza is now known as avozza | 09:55 | |
| *** jerryz has quit IRC | 10:03 | |
| *** salv-orlando has joined #openstack-infra | 10:03 | |
| openstackgerrit | Victor Sergeyev proposed openstack-infra/project-config: Add python3.4 job to Ironic https://review.openstack.org/156530 | 10:05 |
| *** avozza is now known as zz_avozza | 10:06 | |
| *** gulic has joined #openstack-infra | 10:08 | |
| *** pelix has joined #openstack-infra | 10:10 | |
| *** ddieterly has joined #openstack-infra | 10:16 | |
| openstackgerrit | Christian Schwede proposed openstack-infra/lodgeit: Check if MySQL connection are alive https://review.openstack.org/156532 | 10:16 |
| *** otter768 has joined #openstack-infra | 10:16 | |
| *** koolhead_ has quit IRC | 10:17 | |
| openstackgerrit | David Pursehouse proposed openstack-infra/jenkins-job-builder: Fix wrong casing in JDKAxis <name> tag https://review.openstack.org/156519 | 10:17 |
| *** markvoelker has joined #openstack-infra | 10:19 | |
| *** ddieterly has quit IRC | 10:20 | |
| *** otter768 has quit IRC | 10:21 | |
| *** viktors has joined #openstack-infra | 10:21 | |
| *** e0ne is now known as e0ne_ | 10:21 | |
| *** afazekas has joined #openstack-infra | 10:23 | |
| *** markvoelker has quit IRC | 10:24 | |
| *** yfried_ is now known as yfried|afk | 10:24 | |
| *** zz_avozza is now known as avozza | 10:26 | |
| *** e0ne_ has quit IRC | 10:26 | |
| *** mpaolino has joined #openstack-infra | 10:27 | |
| *** mpaolino has quit IRC | 10:27 | |
| *** Guest10052 has quit IRC | 10:29 | |
| *** deepakcs has joined #openstack-infra | 10:30 | |
| *** leopoldj has joined #openstack-infra | 10:32 | |
| *** ilyashakhat_ has quit IRC | 10:32 | |
| *** ilyashakhat has joined #openstack-infra | 10:32 | |
| *** sdake_ has quit IRC | 10:33 | |
| *** sarob has joined #openstack-infra | 10:33 | |
| *** e0ne has joined #openstack-infra | 10:38 | |
| *** leopoldj has quit IRC | 10:38 | |
| *** sarob has quit IRC | 10:41 | |
| *** mpavone has joined #openstack-infra | 10:43 | |
| *** yfried|afk is now known as yfried_ | 10:44 | |
| *** yfried_ is now known as yfried|afk | 10:54 | |
| *** Ala has quit IRC | 10:54 | |
| *** yfried|afk is now known as yfried_ | 10:56 | |
| openstackgerrit | Sergey Lukjanov proposed openstack-infra/puppet-bup: Make backups period configurable https://review.openstack.org/156542 | 10:57 |
| openstackgerrit | yolanda.robla proposed openstack-infra/storyboard: Add an index on accesstokens table tor access_token column https://review.openstack.org/156543 | 10:58 |
| *** pc_m has joined #openstack-infra | 10:59 | |
| *** jcoufal_ has joined #openstack-infra | 11:00 | |
| *** e0ne has quit IRC | 11:01 | |
| *** e0ne has joined #openstack-infra | 11:02 | |
| *** salv-orlando has quit IRC | 11:02 | |
| *** jcoufal has quit IRC | 11:03 | |
| *** rfolco has joined #openstack-infra | 11:10 | |
| *** rlandy has joined #openstack-infra | 11:13 | |
| *** ddieterly has joined #openstack-infra | 11:16 | |
| openstackgerrit | yolanda.robla proposed openstack-infra/nodepool: WIP: Not even close to finished - use shade https://review.openstack.org/156322 | 11:19 |
| openstackgerrit | Sergey Lukjanov proposed openstack-infra/puppet-bup: Make backups period configurable https://review.openstack.org/156542 | 11:19 |
| *** bogdando has quit IRC | 11:20 | |
| *** markvoelker has joined #openstack-infra | 11:20 | |
| *** ddieterly has quit IRC | 11:21 | |
| *** dmsimard has quit IRC | 11:22 | |
| *** bogdando has joined #openstack-infra | 11:23 | |
| *** e0ne has quit IRC | 11:24 | |
| *** e0ne has joined #openstack-infra | 11:25 | |
| *** markvoelker has quit IRC | 11:25 | |
| *** koolhead17 has joined #openstack-infra | 11:27 | |
| *** ilyashak_ has joined #openstack-infra | 11:29 | |
| *** koolhead_ has joined #openstack-infra | 11:29 | |
| *** dmsimard_away has joined #openstack-infra | 11:30 | |
| *** koolhead17 has quit IRC | 11:30 | |
| *** dmsimard_away is now known as dmsimard | 11:30 | |
| *** ashleighfarnham has joined #openstack-infra | 11:31 | |
| *** cdent has joined #openstack-infra | 11:32 | |
| *** dims__ has joined #openstack-infra | 11:32 | |
| *** ilyashak_ has quit IRC | 11:33 | |
| openstackgerrit | Aleksey Ripinen proposed openstack-infra/storyboard: Split auth in api and change user_tokens https://review.openstack.org/153591 | 11:34 |
| *** bogdando has quit IRC | 11:35 | |
| *** ashleighfarnham has quit IRC | 11:36 | |
| *** bogdando has joined #openstack-infra | 11:36 | |
| *** yamamoto has joined #openstack-infra | 11:37 | |
| *** sarob has joined #openstack-infra | 11:38 | |
| openstackgerrit | Sergey Lukjanov proposed openstack-infra/puppet-bup: Make backups period configurable https://review.openstack.org/156542 | 11:42 |
| *** sarob has quit IRC | 11:42 | |
| *** Ala has joined #openstack-infra | 11:44 | |
| *** e0ne is now known as e0ne_ | 11:44 | |
| openstackgerrit | Victor Sergeyev proposed openstack-infra/project-config: Add python3.4 job to Ironic https://review.openstack.org/156530 | 11:48 |
| *** mrmartin has quit IRC | 11:49 | |
| *** e0ne_ has quit IRC | 11:54 | |
| *** e0ne has joined #openstack-infra | 11:55 | |
| *** mrmartin has joined #openstack-infra | 11:56 | |
| *** andreykurilin_ has quit IRC | 11:57 | |
| *** andreykurilin__ has joined #openstack-infra | 11:57 | |
| *** jgallard__ has quit IRC | 11:58 | |
| openstackgerrit | Sergey Lukjanov proposed openstack-infra/puppet-bup: Make backups period configurable https://review.openstack.org/156542 | 11:59 |
| *** salv-orlando has joined #openstack-infra | 12:02 | |
| *** yamamoto has quit IRC | 12:02 | |
| *** yolanda has quit IRC | 12:06 | |
| *** yolanda has joined #openstack-infra | 12:07 | |
| *** deepakcs has quit IRC | 12:10 | |
| *** avozza is now known as zz_avozza | 12:14 | |
| *** otter768 has joined #openstack-infra | 12:17 | |
| *** ddieterly has joined #openstack-infra | 12:17 | |
| *** spredzy is now known as spredzy|afk | 12:17 | |
| *** e0ne is now known as e0ne_ | 12:19 | |
| *** markvoelker has joined #openstack-infra | 12:21 | |
| *** otter768 has quit IRC | 12:22 | |
| *** ddieterly has quit IRC | 12:22 | |
| *** e0ne_ is now known as e0ne | 12:23 | |
| *** dprince has joined #openstack-infra | 12:25 | |
| *** markvoelker has quit IRC | 12:26 | |
| *** BharatK has quit IRC | 12:31 | |
| openstackgerrit | Aleksey Ripinen proposed openstack-infra/storyboard: Datetime instances from SQLAlchemy now all contain timezone. https://review.openstack.org/155083 | 12:32 |
| *** dhritishikhar_ has joined #openstack-infra | 12:33 | |
| openstackgerrit | Aleksey Ripinen proposed openstack-infra/storyboard: Authorization codes validate on TTL. https://review.openstack.org/155084 | 12:37 |
| *** yfried_ is now known as yfried|afk | 12:41 | |
| gilliard | Hi Infra. 'recheck' or 'reverify' ? I get conflicting reports of which I should use. Is there a doc somewhere? | 12:41 |
| *** achuprin_ has quit IRC | 12:41 | |
| *** sarob has joined #openstack-infra | 12:42 | |
| *** antonym has quit IRC | 12:43 | |
| *** yfried|afk is now known as yfried_ | 12:43 | |
| *** LinuxJedi has quit IRC | 12:44 | |
| *** LinuxJedi_ has joined #openstack-infra | 12:44 | |
| *** baoli has joined #openstack-infra | 12:44 | |
| AJaeger | gilliard: recheck, see the infra manual | 12:48 |
| gilliard | I don't see anything in the -dev ML recently, and https://wiki.openstack.org/wiki/GerritJenkinsGit looks out of date | 12:48 |
| AJaeger | gilliard: http://docs.openstack.org/infra/manual/developers.html#automated-testing | 12:48 |
| gilliard | AJaeger: Thanks :) | 12:49 |
| *** baoli has quit IRC | 12:49 | |
| *** mrmartin has quit IRC | 12:49 | |
| *** sarob has quit IRC | 12:49 | |
| *** baoli has joined #openstack-infra | 12:50 | |
| *** salv-orlando has quit IRC | 12:50 | |
| *** salv-orlando has joined #openstack-infra | 12:50 | |
| *** masoz has quit IRC | 12:51 | |
| * AJaeger just added a short note to the wiki page | 12:51 | |
| *** markvoelker has joined #openstack-infra | 12:52 | |
| *** spredzy|afk is now known as spredzy | 12:54 | |
| *** mpavone has quit IRC | 12:54 | |
| *** antonym has joined #openstack-infra | 12:57 | |
| *** yamamoto has joined #openstack-infra | 12:57 | |
| *** zzz_pczesno is now known as pczesno | 13:01 | |
| *** mbacchi has joined #openstack-infra | 13:02 | |
| *** zz_avozza is now known as avozza | 13:05 | |
| openstackgerrit | Aleksey Ripinen proposed openstack-infra/storyboard: Split auth in api and change user_tokens https://review.openstack.org/153591 | 13:06 |
| *** avozza is now known as zz_avozza | 13:07 | |
| *** mpavone has joined #openstack-infra | 13:08 | |
| *** dkranz has quit IRC | 13:11 | |
| *** yamamoto has quit IRC | 13:12 | |
| *** ghostpl_ has quit IRC | 13:13 | |
| *** zz_avozza is now known as avozza | 13:15 | |
| *** mwagner_lap has quit IRC | 13:15 | |
| *** dhritishikhar_ has quit IRC | 13:17 | |
| *** tiswanso has joined #openstack-infra | 13:18 | |
| *** ddieterly has joined #openstack-infra | 13:18 | |
| *** funzo_ is now known as funzo | 13:22 | |
| *** vigneshvar has quit IRC | 13:22 | |
| *** ddieterly has quit IRC | 13:23 | |
| *** weshay has joined #openstack-infra | 13:26 | |
| *** xyang1 has joined #openstack-infra | 13:33 | |
| *** kgiusti has joined #openstack-infra | 13:34 | |
| *** mase_home has joined #openstack-infra | 13:35 | |
| *** yfried_ is now known as yfried|afk | 13:37 | |
| *** Guest10052 has joined #openstack-infra | 13:37 | |
| *** yfried|afk is now known as yfried_ | 13:38 | |
| *** claudiub has joined #openstack-infra | 13:39 | |
| *** koolhead_ has quit IRC | 13:40 | |
| *** afazekas has quit IRC | 13:40 | |
| *** mase_home has quit IRC | 13:41 | |
| *** ghostpl_ has joined #openstack-infra | 13:41 | |
| openstackgerrit | Aleksey Ripinen proposed openstack-infra/storyboard: Added branches to storyboard https://review.openstack.org/150447 | 13:42 |
| *** hdd has joined #openstack-infra | 13:42 | |
| mordred | zaro: when you get in, see the scrollback between gilliard AJaeger and anteaya | 13:44 |
| *** mbacchi has quit IRC | 13:44 | |
| *** mbacchi has joined #openstack-infra | 13:44 | |
| AJaeger | morning, mordred! | 13:44 |
| mordred | morning AJaeger ! | 13:45 |
| *** koolhead17 has joined #openstack-infra | 13:45 | |
| *** otter768 has joined #openstack-infra | 13:45 | |
| *** otter768 has quit IRC | 13:46 | |
| *** amotoki has quit IRC | 13:49 | |
| *** HenryG has quit IRC | 13:49 | |
| *** sarob has joined #openstack-infra | 13:49 | |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul: Add support for layout configuration split https://review.openstack.org/152290 | 13:50 |
| *** mrmartin has joined #openstack-infra | 13:52 | |
| *** HenryG has joined #openstack-infra | 13:52 | |
| *** mbacchi has quit IRC | 13:54 | |
| *** hashar has joined #openstack-infra | 13:54 | |
| *** amotoki has joined #openstack-infra | 13:54 | |
| *** amotoki has quit IRC | 13:54 | |
| *** mbacchi has joined #openstack-infra | 13:54 | |
| *** belmoreira has joined #openstack-infra | 13:54 | |
| *** radez_g0n3 is now known as radez | 13:56 | |
| *** e0ne is now known as e0ne_ | 13:56 | |
| *** dprince has quit IRC | 13:56 | |
| *** sarob has quit IRC | 13:57 | |
| openstackgerrit | Aleksey Ripinen proposed openstack-infra/storyboard: Added milestones https://review.openstack.org/152500 | 13:57 |
| *** e0ne_ is now known as e0ne | 14:00 | |
| *** tiswanso has quit IRC | 14:01 | |
| fungi | amotoki seems to have dropped off, but i added him as the initial member of networking-nec-core and networking-nec-release per scrollback. also made networking-nec-release the owner of networking-nec-ci | 14:01 |
| *** Guest90435 has joined #openstack-infra | 14:02 | |
| *** Guest90435 is now known as annegentle | 14:02 | |
| *** kaisers has joined #openstack-infra | 14:03 | |
| *** avozza is now known as zz_avozza | 14:04 | |
| AJaeger | thanks fungi! | 14:04 |
| AJaeger | and good morning, fungi! | 14:05 |
| fungi | AJaeger: thanks, it's definitely morning here, but the goodness has yet to be empirically determined | 14:06 |
| *** mjturek1 has joined #openstack-infra | 14:07 | |
| *** dustins has joined #openstack-infra | 14:08 | |
| *** signed8bit_ZZZzz is now known as signed8bit | 14:10 | |
| *** signed8bit is now known as signed8bit_ZZZzz | 14:10 | |
| *** signed8bit_ZZZzz is now known as signed8bit | 14:10 | |
| *** changbl has quit IRC | 14:10 | |
| *** bcourt has joined #openstack-infra | 14:10 | |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul: Add support for layout configuration split https://review.openstack.org/152290 | 14:11 |
| *** gordc has joined #openstack-infra | 14:11 | |
| *** ayoung has joined #openstack-infra | 14:12 | |
| AJaeger | fungi: Let's assume it's good until the days convinces us otherwise - instead of the other way round ;) | 14:13 |
| *** rfolco_ has joined #openstack-infra | 14:13 | |
| *** rfolco has quit IRC | 14:13 | |
| *** mpavone has quit IRC | 14:14 | |
| openstackgerrit | gordon chung proposed openstack/requirements: add ceilometermiddleware https://review.openstack.org/154607 | 14:15 |
| *** ddieterly has joined #openstack-infra | 14:15 | |
| *** ddieterly has quit IRC | 14:19 | |
| *** bknudson has joined #openstack-infra | 14:20 | |
| *** tiswanso has joined #openstack-infra | 14:21 | |
| *** tiswanso has quit IRC | 14:22 | |
| *** tiswanso has joined #openstack-infra | 14:22 | |
| *** joesavak has joined #openstack-infra | 14:23 | |
| mordred | AJaeger: I agree | 14:25 |
| *** tiswanso has quit IRC | 14:26 | |
| *** mpavone has joined #openstack-infra | 14:27 | |
| *** tiswanso has joined #openstack-infra | 14:27 | |
| *** pradk has quit IRC | 14:32 | |
| mordred | fungi: btw - it's -8 here and we've gotten probably another 2 inches of snow - so just a normal tuesday | 14:33 |
| mordred | how's the winter storm for you? | 14:33 |
| *** bhunter71 has joined #openstack-infra | 14:33 | |
| *** derekh has quit IRC | 14:33 | |
| *** derekh has joined #openstack-infra | 14:35 | |
| fungi | mordred: mostly ice here--we got freezing rain mixed with sleet. i'm reserving judgement until after lunchtime, but the big issue is that i'm looking at an 80+ mile drive to norfolk on long bridges and poorly-maintained back roads, along with the distinct possibility that after all that the airport might still decide that it can't operate early morning flights tomorrow when everything refreezes | 14:36 |
| fungi | overnight | 14:36 |
| fungi | so it's sort of wait-and-see here still | 14:36 |
| AJaeger | fungi: ;( | 14:37 |
| mordred | fungi: you people with your not-living-inside-of-airports choices | 14:39 |
| *** pradk has joined #openstack-infra | 14:39 | |
| fungi | mordred: those chairs get uncomfortable after a few days | 14:40 |
| *** zz_jgrimm is now known as jgrimm | 14:40 | |
| *** tcammann has quit IRC | 14:40 | |
| *** dalgaaf has joined #openstack-infra | 14:42 | |
| fungi | ttx: anteaya: tristanC: is there any preliminary plan for who's officiating the interim tripleo ptl election? | 14:43 |
| *** mriedem has joined #openstack-infra | 14:45 | |
| *** yamamoto has joined #openstack-infra | 14:46 | |
| *** ddieterly has joined #openstack-infra | 14:46 | |
| tristanC | hum, good question. If no one shows up, should it be up to the TC to find an interim ? | 14:47 |
| fungi | tristanC: i think if nobody working on a particular project wants to be ptl, then nobody's really working on that project and it doesn't need to continue ;) | 14:48 |
| fungi | tristanC: at least in the past, while we've had some close calls, we've never had a situation where nobody self-nominated for an official election | 14:49 |
| openstackgerrit | Sean Dague proposed openstack-infra/devstack-gate: actually save hook logs https://review.openstack.org/156605 | 14:49 |
| sdague | fungi: ^^^ | 14:49 |
| sdague | that would be kind of nice to get in sooner rather than later | 14:50 |
| fungi | tristanC: usually all it takes is jogo threatening to become their ptl, and _someone_ will volunteer | 14:50 |
| *** roaet has joined #openstack-infra | 14:50 | |
| mordred | fungi: what if we elected jogo ptl of everything? | 14:50 |
| *** roaet has left #openstack-infra | 14:51 | |
| *** dkranz has joined #openstack-infra | 14:52 | |
| *** sabeen has joined #openstack-infra | 14:52 | |
| *** hdd has quit IRC | 14:54 | |
| *** sarob has joined #openstack-infra | 14:54 | |
| *** sigmavirus24_awa is now known as sigmavirus24 | 14:54 | |
| sdague | fungi: thanks | 14:55 |
| *** zz_avozza is now known as avozza | 14:56 | |
| *** bcourt has quit IRC | 14:56 | |
| *** bcourt has joined #openstack-infra | 14:57 | |
| *** mtanino has joined #openstack-infra | 14:57 | |
| *** tiswanso has quit IRC | 14:57 | |
| *** wenlock has joined #openstack-infra | 14:58 | |
| fungi | since Ng was the first to reply on that thread, i think he's the defacto nominee | 14:59 |
| Ng | haha | 14:59 |
| Ng | "tag, you're it!" | 15:00 |
| Ng | I hope jslagle stands :) | 15:00 |
| timrc | Hello, for https://review.openstack.org/#/c/156252/ -- would someone who has access to a rax node be willing to test run this for me to confirm it works? | 15:01 |
| *** mwagner_lap has joined #openstack-infra | 15:01 | |
| fungi | timrc: i'll give it a try now | 15:01 |
| *** koolhead17 has quit IRC | 15:01 | |
| *** mestery_ is now known as mestery | 15:01 | |
| timrc | fungi, Thank you! | 15:02 |
| *** mriedem has quit IRC | 15:02 | |
| fungi | timrc: looking at it, i think the first problem you're going to find in rax is lack of a metadata server | 15:02 |
| timrc | fungi, Oh :P lol. | 15:03 |
| fungi | timrc: suspicion confirmed, no response from 169.254.169.254 | 15:03 |
| timrc | fungi, Well I learned something today. | 15:03 |
| *** dims__ has quit IRC | 15:03 | |
| *** cdent_ has joined #openstack-infra | 15:03 | |
| fungi | timrc: however it might be available via configdrive | 15:03 |
| *** cdent has quit IRC | 15:04 | |
| *** cdent_ is now known as cdent | 15:04 | |
| *** dims__ has joined #openstack-infra | 15:04 | |
| *** dims__ has quit IRC | 15:04 | |
| fungi | timrc: do you have a public ssh key i can grant access? i can help you dig into the configdrive details | 15:04 |
| *** mriedem has joined #openstack-infra | 15:04 | |
| *** sarob has quit IRC | 15:05 | |
| *** dims__ has joined #openstack-infra | 15:05 | |
| timrc | fungi, http://paste.openstack.org/show/5zPcyyyJSYnTS0GHAoEV/ | 15:05 |
| mordred | fungi, timrc: honestly, I'd kinda prefer to have nodepool write the UUID like it writes other things | 15:05 |
| timrc | mordred, I thought about that as well. | 15:05 |
| *** mfink_ has joined #openstack-infra | 15:05 | |
| clarkb | I dont think we configdrive yet | 15:05 |
| mordred | metadata service is pretty flaky to be something to be used in a test node | 15:05 |
| clarkb | you have to request configdrive right? | 15:05 |
| mordred | you do | 15:06 |
| fungi | mordred: yeah, we could just stuff it in /etc/nodepool/ | 15:06 |
| *** ghostpl_ has quit IRC | 15:06 | |
| mordred | but - during node boot, you'll have access to the uuid, so writing it to /etc/nodepool would be safe for both clouds and perfectly scalable | 15:06 |
| fungi | via the ready script | 15:06 |
| fungi | clarkb: yeah, i would have to manually boot an instance with configdrive. however i have one handy in another tenant | 15:06 |
| timrc | I'd prefer not to reach out over the network to get this information if that's a possibility (which it sounds like it is) | 15:07 |
| *** e0ne is now known as e0ne_ | 15:07 | |
| clarkb | also why is node uuid importamt in the build log? | 15:07 |
| clarkb | that should be transparent | 15:07 |
| *** amotoki has joined #openstack-infra | 15:07 | |
| clarkb | and if actually needed is retrievable from nodepool logs | 15:08 |
| mordred | clarkb: to help diagnose cloud issues | 15:08 |
| fungi | clarkb: i think it's in case we need to open a case with the provider? | 15:08 |
| clarkb | mordred see nodepool log | 15:08 |
| fungi | but yeah, we do have that data, it's just not published | 15:08 |
| mordred | clarkb: yeah - that's way more not-accessible to people | 15:08 |
| *** e0ne_ is now known as e0ne | 15:08 | |
| clarkb | mordred its accessible to the people that can fix cloud problems or submit tickets | 15:08 |
| mordred | clarkb: that might not be the case in all deployments | 15:09 |
| mordred | clarkb: where, for instance, there may be teams of people who would like to help diagnose and fix underlying cloud issues, but who are not infra-core or infra-root equiv | 15:09 |
| *** joesavak has quit IRC | 15:09 | |
| mordred | also - pretty much anything where the answer is "go look in the log file on the server" should be treated as a bug because it's an instance of bottlenecking working on an issue on us, even upstream | 15:10 |
| mordred | I mean, if the gate is goign bonkers in the middle of the night, one could imagine someone without root looking athte build logs and talking to someone at their cloud-provider company to see if there are endemic cloud problems | 15:11 |
| *** salv-orlando has quit IRC | 15:11 | |
| clarkb | sure, that could happen | 15:11 |
| timrc | clarkb, In which log is the uuid written? nodepool.log or debug.log? | 15:12 |
| clarkb | I dont think its a bad idea was trying to understand the use case behind it | 15:12 |
| clarkb | timrc both | 15:12 |
| clarkb | pretty sure the uuid is logged at info level | 15:12 |
| *** salv-orlando has joined #openstack-infra | 15:13 | |
| timrc | clarkb, Giving people that are having failing runs due to things like network failures the information they need to pass along for someone to diagnose the problem at a physical level. By the time network failures are detected the node is usually deleted, so having that uuid printed to the log at least gives them something to work with. | 15:13 |
| *** ghostpl_ has joined #openstack-infra | 15:13 | |
| openstackgerrit | yolanda.robla proposed openstack-infra/project-config: Initial integration tests for storyboard https://review.openstack.org/156044 | 15:14 |
| timrc | I'm not seeing the uuid in nodepool.log and I found something that I thought was the uuid in the debug log, but it didn't have hyphens.. | 15:14 |
| *** amitgandhinz has joined #openstack-infra | 15:14 | |
| clarkb | timrc I see. It is a good thing if you can get hpcloud to proactively debug things :) | 15:15 |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/system-config: Remove tripleo-test-cloud-hp1 from nodepool config https://review.openstack.org/156615 | 15:15 |
| *** carl_baldwin has joined #openstack-infra | 15:16 | |
| timrc | clarkb, Ah I think my mistake was looking at zuul logs. | 15:17 |
| timrc | And not the more obvious nodepool logs :) | 15:17 |
| mordred | timrc: :) | 15:18 |
| mordred | timrc: and yes, I agree with clarkb - since the network errors you're seeing are almost certainly the same network errors we're seeing - if this helps folks fix it, AWESOME | 15:18 |
| *** rm_work|away is now known as rm_work | 15:18 | |
| anteaya | fungi: we can invoke this: http://git.openstack.org/cgit/openstack/governance/tree/resolutions/20141128-elections-process-for-leaderless-programs.rst | 15:19 |
| timrc | mordred, Okay well, I'll take your advice and write this out to /etc/nodepool and read from that rather than relying on a meta service server being available. | 15:19 |
| *** koolhead17 has joined #openstack-infra | 15:21 | |
| mordred | anteaya: we've had precedent for a PTL stepping down mid-cycle before too - when gabriel stepped down from horizon and david-lyle took over | 15:21 |
| anteaya | yes | 15:21 |
| clarkb | timrc looks like it may be debug for uuid though | 15:21 |
| mordred | anteaya: that time gabriel just appointed david essentially | 15:21 |
| anteaya | the precent for stepping down is appointment | 15:22 |
| anteaya | yes | 15:22 |
| anteaya | in which case no election is needed | 15:22 |
| clarkb | there is an info level launchnode log that logs nodepool id but cloud uuid appears to be debug | 15:22 |
| anteaya | it happened at least one other time as well | 15:22 |
| mordred | it did? | 15:22 |
| anteaya | straight appointment | 15:22 |
| * mordred is losing his memory | 15:22 | |
| anteaya | it might not have been an openstack repo ptl | 15:22 |
| mordred | SpamapS: ^^ | 15:22 |
| anteaya | so many repos | 15:22 |
| anteaya | so the way I see it, 3 opions | 15:23 |
| yolanda | AJaeger, you rock, really, you reviewed the patch almost immediately | 15:23 |
| yolanda | so yes, i turned to non-voting and only for check | 15:23 |
| anteaya | I'm leaving, here is my replacement, nothting to see here, let's get to work | 15:23 |
| anteaya | I'm leaving, peace out, in which case either have an election or invoke the tc resolution | 15:24 |
| yolanda | i need feedback from zuul-cloner, jeblair will take a look today | 15:24 |
| AJaeger | yolanda: fine with me - almost is 6 mins ;) | 15:24 |
| *** timcline has joined #openstack-infra | 15:24 | |
| yolanda | amazing :) | 15:24 |
| *** sdake has joined #openstack-infra | 15:24 | |
| anteaya | so for the election option to work, we need candidates | 15:24 |
| *** jaypipes has quit IRC | 15:24 | |
| AJaeger | so, first a call for candidates, correct? | 15:25 |
| fungi | timrc: just to get you pointed in the right direction, you'll likely want to remotely write something like an /etc/nodepool/node_uuid file in nodepool.NodeLauncher.writeNodepoolInfo() | 15:25 |
| anteaya | if we don't have an appointment yes | 15:26 |
| *** tiswanso has joined #openstack-infra | 15:26 | |
| anteaya | let me read my email now | 15:26 |
| anteaya | then I will come back | 15:26 |
| AJaeger | does anybody know what's HP "changed focus for deployment" is? | 15:26 |
| *** hashar has quit IRC | 15:26 | |
| timrc | fungi, Thanks for the tip. | 15:26 |
| fungi | AJaeger: i would assume there are more details on the way at some point | 15:27 |
| *** naggappan has quit IRC | 15:28 | |
| *** stevemar has joined #openstack-infra | 15:28 | |
| *** yfried_ has quit IRC | 15:28 | |
| *** nelsnelson has joined #openstack-infra | 15:28 | |
| BobBall | Is there any way to automate setting "xenapi_use_agent=false" on an image created by nodepool? | 15:30 |
| mordred | BobBall: it's currently in work | 15:31 |
| clarkb | is that a metadata value? | 15:31 |
| mordred | clarkb: it's a glance property | 15:31 |
| BobBall | mordred: via https://review.openstack.org/#/c/154132/5/tools/update-rax-metadata.sh ? | 15:31 |
| clarkb | nodepool supports arbitrary metadata on images already | 15:31 |
| mordred | BobBall: don't listen to me then - it's already ni | 15:31 |
| mordred | in | 15:31 |
| *** achanda has joined #openstack-infra | 15:32 | |
| BobBall | *looks* | 15:32 |
| sdague | fungi: can you give me a dpkg -l of a base image for nodepool? I'm trying to figure out what installs python-requests | 15:33 |
| sdague | because that seems like devstack doesn't go anywhere near it, and it would be nice to get that out of the base image | 15:33 |
| clarkb | cloud init | 15:33 |
| clarkb | sdague ^ | 15:34 |
| sdague | oh really? | 15:34 |
| rcarrillocruz | hmm | 15:34 |
| *** tiswanso has quit IRC | 15:34 | |
| mordred | it's one of the reasons I want cloud-init out of our base images | 15:34 |
| sdague | hmph | 15:34 |
| rcarrillocruz | folks, do you know if nodepool have support for user-data customization scripts? | 15:34 |
| fungi | clarkb: sdague: interestingly not in rax | 15:34 |
| mordred | fungi: rax doesn't use cloud-init | 15:34 |
| *** tiswanso has joined #openstack-infra | 15:34 | |
| sdague | right, and cloud init needs it because python is stupid pants and doesn't validate certs | 15:35 |
| mordred | yah | 15:35 |
| mordred | so - this is why I've been working on making sure we can have base images that do not need cloud-init or nova-agent | 15:35 |
| mordred | because otherwise we step on test content | 15:35 |
| sdague | ok, I missed that point | 15:35 |
| sdague | good to know | 15:35 |
| openstackgerrit | Steve Martinelli proposed openstack/requirements: Bump pycadf to 0.8.0 https://review.openstack.org/155652 | 15:36 |
| sdague | well, in parallel I started doing this patch - https://review.openstack.org/#/c/156594/ | 15:36 |
| *** zz_dimtruck is now known as dimtruck | 15:36 | |
| sdague | so I think that + your thing gets us close | 15:36 |
| mordred | sdague: AWESOME | 15:36 |
| fungi | yeah, that's been a long time coming | 15:36 |
| mordred | sdague: https://review.openstack.org/#/c/154132/ | 15:36 |
| mordred | there's my thing, fwiw | 15:36 |
| mordred | well, it and the patch before it | 15:36 |
| sdague | but I just realized in the last run that devstack isn't installing python-requests even secondarily | 15:36 |
| mordred | sdague: for now you could put in a hack-around for it to uninstall python-requests | 15:37 |
| *** hdd has joined #openstack-infra | 15:37 | |
| mordred | sdague: it'll cause cloud-init to uninstall, but by that point, that is not a problem | 15:37 |
| sdague | yeh, true | 15:37 |
| clarkb | or just use --upgrade on requests special cased | 15:38 |
| anteaya | fungi: thanks responded | 15:38 |
| sdague | though, in reality, that's also fixable with dtroyer's venv direction as well | 15:38 |
| anteaya | so here is a question, I don't know if we have had a ptl election in the middle of a release before | 15:38 |
| fungi | sdague: have you tried stacking a change on top of that which just dumps dpkg -l into the devstack log and then see what's still missing (and try go get samples from runs in both rax and hpcloud)? | 15:38 |
| anteaya | I'm uncertain what the eligility period is for collecting the electorate | 15:39 |
| anteaya | I believe it should be 365 days prior to the election | 15:39 |
| anteaya | sdague: will there be a tc meeting today? | 15:39 |
| mordred | anteaya: I think ttx actually proposed skipping it today | 15:39 |
| anteaya | sdague: I potentially have a question for the tc | 15:39 |
| anteaya | mordred: yes he did, he left the decision up to sdague | 15:40 |
| anteaya | who would be chairing if one were held | 15:40 |
| sdague | fungi: haven't yet, though I think per mordred's info I'm not sure I need that now | 15:40 |
| sdague | anteaya: oh, I was under the impression that it confirmed that it would be skipped | 15:40 |
| anteaya | I might need to ask the tc about the elegibility period for an impromptu election | 15:40 |
| anteaya | okay | 15:40 |
| *** dkranz has quit IRC | 15:40 | |
| * sdague goes and reads the email more careful | 15:40 | |
| anteaya | well I'm going to go with 365 days then | 15:41 |
| mordred | sdague: that was my impression too | 15:41 |
| *** achanda has quit IRC | 15:41 | |
| sdague | anteaya: given the suddenness on this, I don't think it would be appropriate agenda content anyway, we try to set the agenda at least a day in advance | 15:41 |
| anteaya | sdague: fair enough | 15:42 |
| sdague | so I'd suggest starting an ML thread instead | 15:42 |
| anteaya | true since if we have an election, we have to have nominations open for a week | 15:42 |
| anteaya | so that gives us time to determine the time span for eligible voters | 15:43 |
| anteaya | I'll wait to hear from tripleo if they want to appoint or elect | 15:43 |
| mordred | yeah - some of it's quite moot if there are less than 2 people interested in the positino | 15:43 |
| anteaya | yes | 15:44 |
| anteaya | happy to ferret out the concencous but only if we have at least 2 nominees | 15:44 |
| *** devvesa_ has joined #openstack-infra | 15:44 | |
| *** tiswanso has quit IRC | 15:44 | |
| *** dustins_ has joined #openstack-infra | 15:46 | |
| sdague | ok, I responded to the tc email thread to cancel the meeting | 15:46 |
| *** dustins has quit IRC | 15:46 | |
| *** otter768 has joined #openstack-infra | 15:47 | |
| *** dangers_away is now known as dangers | 15:47 | |
| *** rverchikov has joined #openstack-infra | 15:47 | |
| *** devvesa has quit IRC | 15:47 | |
| *** jgallard__ has joined #openstack-infra | 15:48 | |
| rverchikov | Guys, what is the process of adding users to a group on gerrit? Mostly interested in this group: https://review.openstack.org/#/admin/groups/597,members | 15:48 |
| anteaya | sdague: thanks | 15:48 |
| anteaya | women are present in the channel as well | 15:49 |
| anteaya | some might even be able to respond to your question | 15:49 |
| clarkb | rcarrillocruz: no direct support but nodepool does know how to run a ready script prior to marking a node active | 15:49 |
| clarkb | rcarrillocruz: that can approximate the user data scripts that nova supports | 15:49 |
| rcarrillocruz | yeah, just wondering if the nova thing was already in... for some reason i thought i read something about it in the channel lately, configdrive/user-data or something | 15:50 |
| rcarrillocruz | must have been for shade | 15:50 |
| rcarrillocruz | will search, thx | 15:50 |
| anteaya | mordred: it was the marconi/zaqar appointment | 15:51 |
| *** otter768 has quit IRC | 15:51 | |
| clarkb | rcarrillocruz: config drive is independent of the user data here since I don't think nodepool knows how to supply user data to the config drive that is eventually mounted | 15:54 |
| *** r-daneel has joined #openstack-infra | 15:54 | |
| *** dmsimard is now known as dmsimard_away | 15:54 | |
| rverchikov | anteaya: I didn't want to be offensive, it's just my poor english, to me "guys" means "all", nothing personal | 15:56 |
| anteaya | all works then | 15:56 |
| *** marun has joined #openstack-infra | 15:57 | |
| anteaya | you ask fungi to add you to the cloudv-core group in gerrit | 15:57 |
| *** dkranz has joined #openstack-infra | 15:57 | |
| anteaya | if he needs more info he will ask | 15:57 |
| anteaya | for instance is rverchikov your gerrit username? | 15:58 |
| rverchikov | yes | 15:58 |
| anteaya | then that makes it easy for him | 15:58 |
| clarkb | any gerrit admin can do it... | 15:58 |
| anteaya | yes | 15:58 |
| anteaya | clarkb: able to do it? | 15:58 |
| clarkb | ya taking a look now | 15:58 |
| anteaya | clarkb: thanks | 15:59 |
| rverchikov | clarkb: Thanks! | 15:59 |
| *** sdague has quit IRC | 16:00 | |
| sdake_ws | can someone with magic powers use #endmeeting in openstack-meeting-alt? | 16:01 |
| sdake_ws | someone left theirmeeting open | 16:01 |
| *** tiswanso has joined #openstack-infra | 16:01 | |
| *** sarob has joined #openstack-infra | 16:01 | |
| *** baoli has quit IRC | 16:01 | |
| *** marun has quit IRC | 16:01 | |
| sdake_ws | nm, all set | 16:01 |
| *** sdague has joined #openstack-infra | 16:02 | |
| fungi | sdake_ws: if it's been more than an hour, anyone can | 16:02 |
| clarkb | sdake_ws: fwiw anyone can do it an hour after the meeting was started | 16:02 |
| rverchikov | anteaya, clarkb: could you share a link which describes the process of adding new users to a gerrit group? I think I'm not the only person who should be added to cloudv-core group (and I don't have a list of all users at the moment) | 16:03 |
| anteaya | rverchikov: we add one, you add the rest yourself | 16:04 |
| clarkb | rverchikov: you have been added to cloudv-core and cloudv-release | 16:04 |
| rverchikov | great, thank you! | 16:04 |
| anteaya | rverchikov: go to the core page you should see a text box and a button named add | 16:04 |
| anteaya | add whoever you need to | 16:04 |
| *** achanda has joined #openstack-infra | 16:05 | |
| rverchikov | yes, seem to be working, thanks! | 16:05 |
| sdake_ws | fungi clarkb ook thanks I tried #endmeeting and it didn't seem to work | 16:06 |
| *** achanda has quit IRC | 16:06 | |
| *** nfedotov1 has quit IRC | 16:07 | |
| anteaya | sdake_ws: sc68cal ended it for you | 16:07 |
| anteaya | at 16:01 | 16:07 |
| *** sarob has quit IRC | 16:08 | |
| *** erikmwil_ has joined #openstack-infra | 16:08 | |
| *** erikmwilson is now known as Guest10876 | 16:08 | |
| *** erikmwil_ is now known as erikmwilson | 16:08 | |
| *** mriedem has quit IRC | 16:08 | |
| *** mriedem1 has joined #openstack-infra | 16:08 | |
| *** erikmwil- has joined #openstack-infra | 16:08 | |
| *** rverchikov has quit IRC | 16:08 | |
| SergeyLukjanov | folks, could you please take a look on my attempt on puppet scripts? https://review.openstack.org/#/c/156542/ (make bup backup period configurable externally) | 16:08 |
| sc68cal | sorry I got interrupted for a couple seconds and didn't end the meeting at 1600 sharp | 16:09 |
| anteaya | sdake_ws: it didn't work for you since the meeting was still in progress and had to be ended by the chair | 16:09 |
| *** e0ne has quit IRC | 16:09 | |
| *** spredzy is now known as spredzy|afk | 16:09 | |
| sdake | ya its al lgood folks | 16:10 |
| *** armax has joined #openstack-infra | 16:11 | |
| *** reed has joined #openstack-infra | 16:11 | |
| *** mpavone has quit IRC | 16:13 | |
| *** amotoki has quit IRC | 16:13 | |
| *** mfink__ has joined #openstack-infra | 16:13 | |
| *** mfink_ has quit IRC | 16:14 | |
| *** skolekonov has quit IRC | 16:14 | |
| sdague | so... the ironic jobs are all fail now? - http://logs.openstack.org/05/156605/1/check/check-tempest-dsvm-ironic-pxe_ssh/7a823fa//logs/devstacklog.txt.gz ? | 16:15 |
| sdague | that's the 2nd change of mine today that it's errored like that | 16:15 |
| *** mfink_ has joined #openstack-infra | 16:17 | |
| clarkb | possible they added a new image and we haevn't cached it yet? | 16:17 |
| clarkb | SergeyLukjanov: lgtm | 16:17 |
| *** mfink__ has quit IRC | 16:17 | |
| clarkb | sdague: I don't see any new images in a quick grep through devstack's git log | 16:18 |
| SergeyLukjanov | clarkb, thx | 16:19 |
| *** arxcruz has quit IRC | 16:19 | |
| openstackgerrit | Merged openstack-infra/project-config: Use remote git path in DEVSTACK_LOCAL_CONFIG https://review.openstack.org/155217 | 16:19 |
| clarkb | sdague: oh I see that part of the job is running DIB to make an image | 16:19 |
| clarkb | sdague: and that image builds needs the trusty base image. So likely just missed caching here | 16:19 |
| *** gulic has quit IRC | 16:20 | |
| dvorak | I'm probably missing something obvious, but I've got gate tests failing due to rubygems being flakey. How do you retrigger gate tests to run again? | 16:21 |
| dvorak | it looks like "recheck no bug" just retriggers the check tests | 16:22 |
| clarkb | dvorak: `recheck` and `reverify` are now equivalent and you have to pass check before going to the gate. So if you have a -1 then recheck should send you to check to get a +1 and if ou get a +1 it should go to the gate | 16:22 |
| dvorak | huh, ok, so the gate tests may just be queued and not run yet? | 16:23 |
| clarkb | dvorak: maybe you can check at http://status.openstack.org/zuul/ | 16:23 |
| AJaeger | dvorak: just "recheck" is enough, no need to pass "no bug" | 16:23 |
| dvorak | ok. I'll do some more research, thanks :) | 16:23 |
| *** ilyashak_ has joined #openstack-infra | 16:26 | |
| *** changbl has joined #openstack-infra | 16:26 | |
| openstackgerrit | yolanda.robla proposed openstack-infra/nodepool: WIP: Not even close to finished - use shade https://review.openstack.org/156322 | 16:26 |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul: Add support for layout configuration split https://review.openstack.org/152290 | 16:27 |
| *** abhirc has quit IRC | 16:28 | |
| *** dizquierdo has quit IRC | 16:29 | |
| *** dkliban_afk is now known as dkliban | 16:29 | |
| *** andreykurilin__ has quit IRC | 16:34 | |
| *** ddieterly has quit IRC | 16:34 | |
| *** andreykurilin_ has joined #openstack-infra | 16:34 | |
| sdague | clarkb: so nothing has changed recently in devstack in this area | 16:34 |
| sdague | what specifies that base image? | 16:35 |
| clarkb | sdague: disk image builder | 16:36 |
| clarkb | BobBall: I have reviewed your nodepool change to enable config drive as an option. couple of small things noted inline | 16:36 |
| sdague | so disk image builder had a release, and now the ironic jobs are the fail until some time in the future? | 16:36 |
| clarkb | sdague: possibly, or ironic changed their base image from say precise to trusty | 16:37 |
| clarkb | sdague: OR if you look at the log the host the file lives on doesn't respond so it could be something changed ubuntu side | 16:37 |
| *** hdd has quit IRC | 16:37 | |
| *** derekh has quit IRC | 16:37 | |
| *** sputnik13 has quit IRC | 16:38 | |
| clarkb | I guess I can check if it pings from here | 16:38 |
| *** derekh has joined #openstack-infra | 16:38 | |
| clarkb | ya `telnet 91.189.88.141 80` establishes a connection from here | 16:38 |
| BobBall | thanks clarkb - new patchset on it's way! | 16:39 |
| * BobBall just does a few tests first... | 16:39 | |
| sdague | so that to me indicates that this whole test path is too fragile to be voting | 16:39 |
| *** sputnik13 has joined #openstack-infra | 16:40 | |
| mordred | sdague: if we turns off voting on a piece of the stack every time a part of the environment caused a problem, there would be no gating jobs in openstack | 16:40 |
| sdague | mordred: well, my patience with jobs connecting to the open internet is low | 16:42 |
| jeblair | clarkb: nodepool image log mentions trusty-server-cloudimg-amd64-root.tar.gz | 16:42 |
| jeblair | sdague: ^ | 16:42 |
| *** ilyashak_ has quit IRC | 16:42 | |
| *** sputnik13 has quit IRC | 16:42 | |
| clarkb | jeblair: ya I think devstack knows where to get that image but dib doesn't | 16:42 |
| jeblair | clarkb: that sounds reasonable | 16:42 |
| mordred | sdague: sure. but it seems to be a bug in the script, not a "whole test path too fragile" | 16:43 |
| jeblair | at any rate, way forward seems to be to have that script use the local cache... | 16:44 |
| mordred | yup | 16:44 |
| *** esker has joined #openstack-infra | 16:44 | |
| sdague | ok, so who's working on it? | 16:44 |
| mordred | greghaynes, SpamapS: ^^ you're both smart in this area | 16:44 |
| *** dtantsur is now known as dtantsur|afk | 16:44 | |
| jeblair | devananda: ^ | 16:44 |
| sdague | more importantly, dib should have a fail if talks to the network mode | 16:44 |
| mordred | oh yeah. devananda too :) | 16:44 |
| mordred | sdague: ++ | 16:45 |
| jeblair | that would be swell | 16:45 |
| sdague | but, the fact that it doesn't, does mean it's adding a ton of fragility, on a job that has managed to stick itself on a ton of projects | 16:45 |
| sdague | jeblair: unrelated, but on a patch that got nailed by this, https://review.openstack.org/#/c/156605/ | 16:46 |
| sdague | that no longer passes unit tests, because the unit tests now try to sudo | 16:46 |
| sdague | got thoughts about the way to resolve that? | 16:46 |
| openstackgerrit | Andreas Jaeger proposed openstack/requirements: Remove failing project nova-docker https://review.openstack.org/156260 | 16:46 |
| * greghaynes reads backscroll | 16:46 | |
| mordred | greghaynes: tl;dr - a) ironic's use of dib is not making use of the pre-downloaded base image that devstack-gate provides | 16:47 |
| *** devanand_ has joined #openstack-infra | 16:48 | |
| mordred | greghaynes: b) we need to make sure dib has as "fail on attempt to access network" mode | 16:48 |
| mordred | greghaynes: c) as a follow on from that, we really do need better integration between dib and d-g pre-caching | 16:48 |
| clarkb | sdague: should those test sudo? | 16:48 |
| sdague | clarkb: the test does not | 16:48 |
| sdague | directly | 16:48 |
| mordred | greghaynes: in descending order of immediate importance and ascending order probably of difficulty | 16:48 |
| greghaynes | There is an "offline mode" for dib, although its possible theres some places that dont use the setting properly | 16:48 |
| sdague | it calls a function which needs sudo | 16:48 |
| clarkb | sdague: I see its the function, ya | 16:48 |
| *** ihrachyshka has quit IRC | 16:49 | |
| *** andreykurilin_ has quit IRC | 16:49 | |
| sdague | basically we've not been collecting hook logs for some large amount of time | 16:49 |
| mordred | greghaynes: place that don't use the setting properly are just bugs - so the existence of the setting is awesome | 16:49 |
| openstackgerrit | yolanda.robla proposed openstack-infra/nodepool: WIP: Not even close to finished - use shade https://review.openstack.org/156322 | 16:49 |
| greghaynes | Yep | 16:49 |
| sdague | because the collector didn't have permissions | 16:49 |
| jeblair | sdague: can we replace it with a fake? | 16:49 |
| *** andreykurilin_ has joined #openstack-infra | 16:49 | |
| jeblair | sdague: put a 'sudo' in the start of $PATH? | 16:49 |
| clarkb | sdague: we could remove the sudo check from the job, though if there is some way to avoid that it would be nice (I like unittests that don't require priveleges) | 16:49 |
| clarkb | jeblair: ++ | 16:49 |
| jeblair | clarkb: yeah, i'd rather be able to run unit tests locally | 16:50 |
| sdague | oh, sure, pathing trick sounds interesting | 16:50 |
| clarkb | would a sudo function work too? | 16:50 |
| clarkb | might be easier in the context of the unittests | 16:50 |
| jeblair | or also call "$SUDO" and then define that conditionally... | 16:50 |
| *** bknudson has quit IRC | 16:50 | |
| *** viktors is now known as viktors|afk | 16:51 | |
| *** coolsvap is now known as coolsvap_ | 16:51 | |
| jeblair | clarkb, sdague: yeah, actually i think either the sudo function or variable approaches are probably preferable to pathing | 16:51 |
| openstackgerrit | Matt Riedemann proposed openstack-infra/elastic-recheck: Add query for ironic grenade bug 1422520 https://review.openstack.org/156453 | 16:51 |
| openstack | bug 1422520 in grenade "check-grenade-dsvm-ironic-sideways failing test_list_baremetal_node" [Undecided,New] https://launchpad.net/bugs/1422520 | 16:51 |
| *** devvesa_ has quit IRC | 16:52 | |
| *** rwsu has joined #openstack-infra | 16:52 | |
| greghaynes | Yep, looks like cache-url doesnt make use of DIB_OFFLINE like it should | 16:52 |
| greghaynes | so thats a bug | 16:52 |
| *** markus_z has quit IRC | 16:53 | |
| *** sarob has joined #openstack-infra | 16:53 | |
| greghaynes | and also looks like the offline mode isnt used in that test | 16:54 |
| *** hdd has joined #openstack-infra | 16:55 | |
| sdague | greghaynes: could we get some dib tests that try to build a real image as part of that and tests some of these assumptions. It looks like the first time we try to do any of this for real is the check-tempest-dsvm-ironic-pxe_ssh job | 16:56 |
| openstackgerrit | Sean Dague proposed openstack-infra/devstack-gate: actually save hook logs https://review.openstack.org/156605 | 16:57 |
| sdague | jeblair: so I made the minimum possible change for that, I think adding additional SUDO use can be future patches | 16:58 |
| mordred | yeah - if we could have a job that gates dib on building an image in offline mode that contains the elements that the ironic job would be containing (minus any elements from ironic)- that would be neat | 16:58 |
| openstackgerrit | Merged openstack-infra/elastic-recheck: Add query for ironic grenade bug 1422520 https://review.openstack.org/156453 | 16:58 |
| openstack | bug 1422520 in grenade "check-grenade-dsvm-ironic-sideways failing test_list_baremetal_node" [Undecided,In progress] https://launchpad.net/bugs/1422520 - Assigned to Adam Gandelman (gandelman-a) | 16:58 |
| mordred | and would give us a template for a job that we could use to test the combination of dib elements that infra would be using too | 16:58 |
| mordred | ooh | 16:58 |
| greghaynes | Sure, checking that offline mode 'works' is hard by dib itself since elements can do whatever they want. In this case it would just pass even if you passed offline mode because the element just doesnt respect it | 16:59 |
| jeblair | sdague: cool, thx | 16:59 |
| *** spredzy|afk is now known as spredzy | 16:59 | |
| mordred | greghaynes: I betcha that ubuntu-minimal does not respect it all, come to think of it | 16:59 |
| sdague | greghaynes: what tools does it use to call out to the network? | 16:59 |
| mordred | greghaynes: how hard would it be to run dib in a container without outbound network access? | 16:59 |
| sdague | you could poison them all | 16:59 |
| mordred | sdague: anything - elements are comprised of executable scripts | 17:00 |
| greghaynes | curl, git and pip are the big ones | 17:00 |
| *** scheuran has quit IRC | 17:00 | |
| *** doug-fish has joined #openstack-infra | 17:00 | |
| greghaynes | or just firewall off outbound port 80, 443 would hit most of it | 17:00 |
| mordred | greghaynes: and apt-get and yum | 17:00 |
| greghaynes | oh right | 17:00 |
| mordred | greghaynes: ++ | 17:00 |
| greghaynes | that | 17:00 |
| jeblair | or put a counter on those ports | 17:00 |
| greghaynes | ooo | 17:01 |
| mordred | ++ | 17:01 |
| greghaynes | so yes, sounds like there are solutions :) | 17:01 |
| *** Guest31726 is now known as dank_ | 17:01 | |
| sdague | greghaynes: you want to take https://bugs.launchpad.net/openstack-ci/+bug/1375488 and assign it to dib as well for this? Because I think those fixes you are working on is the solution | 17:02 |
| openstack | Launchpad bug 1375488 in OpenStack-Gate "ironic devstack jobs downloading from cloud-images.ubuntu.com occasionally fails" [Undecided,New] | 17:02 |
| greghaynes | Yep | 17:03 |
| devanand_ | I think I'd prefer if the ironic jobs could just consume artifacts built by previous jobs, rather than running dib. | 17:04 |
| *** claudiub_ has joined #openstack-infra | 17:04 | |
| *** marun has joined #openstack-infra | 17:04 | |
| sdague | devanand_: well no such other jobs exist at this point, right? | 17:04 |
| openstackgerrit | Merged openstack-infra/infra-specs: Update the test-metrics-db spec to reflect recent changes https://review.openstack.org/141855 | 17:04 |
| mtreinish | ^^^ heh, I had forgotten about that | 17:05 |
| devanand_ | Also, I need to make irssi/irccloud use the same account | 17:05 |
| *** baoli has joined #openstack-infra | 17:05 | |
| devanand_ | sdague: the coreos image builder used by IPA | 17:05 |
| sdague | devanand_: hey, if you want to fix the jobs to not be doing this on the fly, more power to you :) | 17:06 |
| *** jcoufal_ has quit IRC | 17:06 | |
| clarkb | yolanda: commented on the storyboard integration tests change. let me know if you need more clarification on my comments | 17:06 |
| devanand_ | jroll: the agent job fetches the image, rather than builds it, right? | 17:06 |
| yolanda | ok, let me take a look | 17:06 |
| *** ivar-lazzaro has joined #openstack-infra | 17:06 | |
| zaro | morning | 17:06 |
| sdague | I mostly want whatever makes this more solid so that things like fixing devstack-gate bugs aren't blocked by it | 17:07 |
| jroll | devanand_: agent_ssh fetches the image, agent_ssh-src builds it. run on ironic and ipa respectively | 17:07 |
| *** claudiub has quit IRC | 17:07 | |
| devanand_ | sdague: totally agree. I'm suggesting a better approach, which is what we're using with IPA | 17:07 |
| yolanda | clarkb, so there is no need of specifying master branch? i want that zuul-cloner runs the clone on zuul_ref for the {name} project, and the master branch for the {target_name} project | 17:07 |
| yolanda | will zuul-cloner figure itself? | 17:07 |
| yolanda | it woud be much simpler then | 17:08 |
| yolanda | as i can create same jobs for the 2 projects | 17:08 |
| jroll | devanand_: an IPA post job builds the image and publishes it, I think dib could do the same | 17:08 |
| clarkb | yolanda: zuul-cloner knows how to read the ZUUL_* env vars to checkout the correct branches based on what zuul has told it | 17:08 |
| jroll | sdague: ^ | 17:08 |
| yolanda | oh, that's amazin | 17:08 |
| yolanda | amazing | 17:08 |
| yolanda | so just 2 zuul-cloners without caring of branches and will do the job, right? | 17:09 |
| *** ivar-lazzaro has quit IRC | 17:09 | |
| *** dhritishikhar_ has joined #openstack-infra | 17:09 | |
| clarkb | yolanda: just one zuul-cloner | 17:09 |
| yolanda | but i need 2 projects to be cloned | 17:09 |
| clarkb | yolanda: with a single list of projects to checkout | 17:09 |
| yolanda | ah ok | 17:09 |
| clarkb | yolanda: ya you provide a list | 17:09 |
| devanand_ | greghaynes: ^ | 17:09 |
| *** dkranz has quit IRC | 17:10 | |
| *** ivar-lazzaro has joined #openstack-infra | 17:10 | |
| clarkb | yolanda: let me find an example | 17:10 |
| yolanda | i digged on infra, but only saw the infra-config, that was one repo | 17:10 |
| *** emagana_ has quit IRC | 17:10 | |
| clarkb | yolanda: https://git.openstack.org/cgit/openstack-infra/system-config/tree/tools/apply-test.sh#n44 starts about there | 17:10 |
| *** tim_o has joined #openstack-infra | 17:11 | |
| clarkb | yolanda: I don't think the clonemap.yaml is necessary for your use case so you should be able to ignore it | 17:11 |
| *** Ala has quit IRC | 17:11 | |
| yolanda | so either i use the clonemap and specify the list of projects there, or i pass them from command line, right? | 17:11 |
| clarkb | yolanda: no you always have to pass in from the command line I think. The clone map just overrides where things are cloned to if you need that | 17:12 |
| yolanda | not my case then | 17:12 |
| clarkb | yolanda: since puppet expects repos in specific locations we use the clonemap in that job but I don't think your job needs it | 17:12 |
| yolanda | clarkb, that's a really useful tool, i regret not having looked at that before | 17:13 |
| yolanda | very useful for some of our downstream builds | 17:13 |
| *** changbl has quit IRC | 17:13 | |
| yolanda | clarkb, to use it , we need to have that zuul binary installed on the slaves, right? | 17:13 |
| yolanda | i think we are behind that | 17:13 |
| *** ddieterly has joined #openstack-infra | 17:14 | |
| *** ddieterly has quit IRC | 17:14 | |
| *** ddieterly has joined #openstack-infra | 17:15 | |
| clarkb | yolanda: yes, our image builds put it in a virtualenv then put that virtualenv's zuul-cloner executable in $PATH | 17:15 |
| yolanda | it's worth to take a look, really | 17:16 |
| *** unicell has joined #openstack-infra | 17:19 | |
| greghaynes | devanand_: hrm? the coreos-image-builder comment? | 17:19 |
| *** dmsimard_away is now known as dmsimard | 17:19 | |
| *** ashleighfarnham has joined #openstack-infra | 17:20 | |
| *** ashleighfarnham has quit IRC | 17:20 | |
| *** ashleighfarnham has joined #openstack-infra | 17:21 | |
| *** belmoreira has quit IRC | 17:21 | |
| greghaynes | Im guessing that must also have some type of pre-populated cache tooling, otherwise it has the same issue... | 17:21 |
| *** doude has quit IRC | 17:21 | |
| *** jgallard__ has quit IRC | 17:22 | |
| *** carl_baldwin has quit IRC | 17:23 | |
| *** jistr has quit IRC | 17:24 | |
| sdague | greghaynes: I bet it doesn't, but it's a post job, so when it fails it just fails and doesn't stop anything | 17:25 |
| *** dkranz has joined #openstack-infra | 17:26 | |
| *** jaypipes has joined #openstack-infra | 17:26 | |
| stevemar | sdague, dhellmann SergeyLukjanov could i get a review of https://review.openstack.org/#/c/155652/ as it's holding up a keystone bp from landing in k3 | 17:27 |
| stevemar | morganfainberg, ^ | 17:27 |
| openstackgerrit | Merged openstack-infra/project-config: Add test job for sheepdog devstack plugin https://review.openstack.org/154605 | 17:29 |
| *** annegent_ has joined #openstack-infra | 17:29 | |
| *** wenlock has quit IRC | 17:30 | |
| *** bdpayne has joined #openstack-infra | 17:30 | |
| zaro | nibalizer: replied to your concern on https://review.openstack.org/#/c/155142/ | 17:30 |
| nibalizer | woot | 17:31 |
| *** tkelsey has joined #openstack-infra | 17:31 | |
| *** asselin has quit IRC | 17:32 | |
| *** asselin has joined #openstack-infra | 17:32 | |
| sdague | stevemar: why is that holding things up? | 17:32 |
| sdague | oh, I see | 17:33 |
| openstackgerrit | yolanda.robla proposed openstack-infra/nodepool: WIP: Not even close to finished - use shade https://review.openstack.org/156322 | 17:33 |
| *** miqui_ has quit IRC | 17:34 | |
| openstackgerrit | Sean Dague proposed openstack-infra/project-config: have swift functional tests use tox https://review.openstack.org/156676 | 17:35 |
| *** patrickeast has joined #openstack-infra | 17:36 | |
| sdague | notmyname: can you take a look at https://review.openstack.org/156676 when you get about, I think that should be a safe flip for swift | 17:36 |
| *** miqui has quit IRC | 17:37 | |
| notmyname | sdague: ya, I just did. just trying to grok it now | 17:37 |
| sdague | mordred: also, ^^^ is needed to be able to land the devstack package pruning | 17:37 |
| notmyname | sdague: so using tox instead of the included .functests script? | 17:37 |
| sdague | notmyname: correct | 17:37 |
| sdague | which means the test requirements will get installed for the test | 17:37 |
| *** rverchikov has joined #openstack-infra | 17:37 | |
| notmyname | sdague: I don't understand why that changes requirements | 17:37 |
| *** e0ne has joined #openstack-infra | 17:38 | |
| sdague | so, if you tox -e func | 17:38 |
| *** asselin has quit IRC | 17:38 | |
| sdague | it builds a virtenv with requirements.txt and test-requirements.txt automatically | 17:38 |
| sdague | and uses that | 17:38 |
| notmyname | sdague: instead of just whatever is installed on the system | 17:38 |
| notmyname | gotcha | 17:38 |
| sdague | yep | 17:38 |
| openstackgerrit | Merged openstack-infra/project-config: Remove nova v3 regex from devstack-virtual-ironic https://review.openstack.org/155564 | 17:38 |
| clarkb | sdague: do you have a plan for mysql-python, libvirt-python and friends yet? | 17:39 |
| notmyname | sdague: I'm a huge fan of using project-specific requirements (if they don't conflict with global requirements) for testing! | 17:39 |
| clarkb | sdague: mostly asking because there is a spec in oslo to deal with optional deps (and figure these things are related) | 17:39 |
| *** ghostpl_ has quit IRC | 17:39 | |
| *** asselin has joined #openstack-infra | 17:39 | |
| sdague | clarkb: not yet, I'm just trying to delete things like python-nose and python-mock from the devstack package list | 17:39 |
| clarkb | ++ | 17:40 |
| sdague | and once swift is using a tox target for functional tests, we should be able to, as it will install what it needs on demand | 17:40 |
| openstackgerrit | Marton Kiss proposed openstack-infra/infra-specs: Add migrate askbot to infra spec https://review.openstack.org/154061 | 17:41 |
| dhellmann | stevemar, sdague : it's nice that an unintended consequence of capping requirements in stable forces us to update minimums when we need something new :-) | 17:44 |
| sdague | dhellmann: yeh, though honestly it would also be cool if our code had sane fallbacks :) | 17:45 |
| dhellmann | sdague: one step at a time :-) | 17:45 |
| *** koolhead17 has quit IRC | 17:45 | |
| dhellmann | at least this way the question is raised | 17:45 |
| sdague | agreed, I did +2 the patch :) | 17:45 |
| clarkb | sdague: wold really help upgrade flexibility to have that | 17:45 |
| *** cipcosma has quit IRC | 17:45 | |
| dhellmann | sdague: I saw, I just approved it | 17:45 |
| sdague | clarkb: yeh, it opens a whole other ball of wax | 17:45 |
| * dhellmann thinks maybe the logo for openstack should be someone juggling | 17:46 | |
| *** tqtran has joined #openstack-infra | 17:46 | |
| *** thedodd has joined #openstack-infra | 17:46 | |
| *** hashar has joined #openstack-infra | 17:47 | |
| jeblair | marun: question on neutron-functional -- does that job exercise any openstack components other than neutron? eg, nova? | 17:47 |
| *** otter768 has joined #openstack-infra | 17:47 | |
| marun | jeblair: it does not | 17:48 |
| marun | jeblair: the intention is neutron-only in-tree. We'll use fakes if we need something nova-like. | 17:48 |
| jeblair | marun: sounds good then, thanks | 17:48 |
| openstackgerrit | Max Rasskazov proposed openstack-infra/jenkins-job-builder: Add support of specify of multiple ssh-agent credentials https://review.openstack.org/156681 | 17:48 |
| jeblair | marun: (just needed full context to understand the devstack-gate $PROJECTS change and related work) | 17:49 |
| zaro | gilliard, anteaya, AJaeger, mordred : missing '_' doesn't seem to be an issue in Gerrit 2.9, here's a test: https://review-dev.openstack.org/#/c/5331/1..2/driver.py | 17:49 |
| clarkb | jeblair: 156222 does not have any of the potential issues that the skip-if change(s) had right? safe for me to approve? | 17:49 |
| jeblair | clarkb: i think that should be fine | 17:49 |
| marun | jeblair: cool | 17:49 |
| *** BobBall is now known as BobBall_AWOL | 17:49 | |
| anteaya | zaro: I think it was the new view that caused the problem | 17:50 |
| openstackgerrit | Max Rasskazov proposed openstack-infra/jenkins-job-builder: Add support of multiple ssh-agent credentials https://review.openstack.org/156681 | 17:50 |
| *** cipcosma has joined #openstack-infra | 17:50 | |
| jeblair | marun: are you planning on hacking on zuul to make the metajob booleans trinary? | 17:50 |
| notmyname | sdague: I left my +1 on the review. and I mentioned it in -swift so that other people are aware | 17:50 |
| notmyname | sdague: thanks | 17:50 |
| *** MaxV has quit IRC | 17:50 | |
| zaro | anteaya: yep, i checked with new screen and i don't see the problem. you can verify yourself. | 17:50 |
| marun | jeblair: I'm working on it, yes. Will have a patch up today. | 17:50 |
| fungi | clarkb: we're still booting devstack-trusty-dib nodes in hpcloud... any reason? should we clean that up now? | 17:50 |
| anteaya | zaro: trying now | 17:50 |
| marun | jeblair: it looks like a trivial fix, but I'm debugging seeminly unrelated test failures. | 17:50 |
| clarkb | fungi: at the time I thought I would use them with rax | 17:51 |
| fungi | ahh, okay | 17:51 |
| clarkb | fungi: but it has been slow going. we can probably remove them now and add something similar back in when rax is ready | 17:51 |
| fungi | well, i just held one out of convenience because ready devstack-trusty nodes start running jobs before i can grab them in jenkins and tell them not to | 17:51 |
| *** carl_baldwin has joined #openstack-infra | 17:52 | |
| jeblair | marun: cool; so for sanity, we probably want to hold further skip-ifs on that (otherwise, we'll have to have extra-rigid rules about the order they appear in the file (as opposed to just the normal rules)) | 17:52 |
| marun | jeblair: definitely a good idea | 17:52 |
| *** gyee has joined #openstack-infra | 17:52 | |
| *** otter768 has quit IRC | 17:52 | |
| jeblair | marun: feel free to push up a wip if you want me to look at any test failures | 17:52 |
| anteaya | zaro: yes I see the underscores in line 5346: https://review-dev.openstack.org/#/c/5331/2/driver.py | 17:52 |
| anteaya | zaro: thanks | 17:52 |
| marun | jeblair: will do. I can push up what I have sans-tests shortly. | 17:53 |
| *** hashar has quit IRC | 17:53 | |
| *** vilobhmm has joined #openstack-infra | 17:55 | |
| *** carl_baldwin_ has joined #openstack-infra | 17:55 | |
| *** bhuvan_ has joined #openstack-infra | 17:56 | |
| *** carl_baldwin has quit IRC | 17:56 | |
| *** carl_baldwin_ is now known as carl_baldwin | 17:56 | |
| *** gulic has joined #openstack-infra | 17:58 | |
| *** yamamoto has quit IRC | 17:59 | |
| openstackgerrit | Merged openstack-infra/project-config: Run translation jobs only if translation files change https://review.openstack.org/156222 | 17:59 |
| *** bknudson has joined #openstack-infra | 18:01 | |
| *** rverchikov has quit IRC | 18:02 | |
| *** rverchikov has joined #openstack-infra | 18:02 | |
| *** derekh has quit IRC | 18:02 | |
| *** AJaeger has quit IRC | 18:03 | |
| sdague | mordred can you take a look at - https://review.openstack.org/#/c/156356/1 for correctness in building and using wheels in devstack? | 18:04 |
| *** abhirc has joined #openstack-infra | 18:05 | |
| *** AJaeger has joined #openstack-infra | 18:06 | |
| *** yamahata has joined #openstack-infra | 18:06 | |
| *** davideagnello has joined #openstack-infra | 18:09 | |
| *** ZZelle_ has joined #openstack-infra | 18:09 | |
| clarkb | sdague: I am not sure WHEELHOUSE as an env var does the right thing, trying to confirm | 18:10 |
| clarkb | sdague: the command line arg to pip is --wheel-dir which means I think PIP_WHEEL_DIR is what you want? | 18:10 |
| *** harlowja_away is now known as harlowja_ | 18:10 | |
| openstackgerrit | Khai Do proposed openstack-infra/project-config: Use job templates for gerritlib builds. https://review.openstack.org/156068 | 18:11 |
| greghaynes | clarkb: I think that sets PIP_WHEEL_DIR to WHEELHOUSE | 18:11 |
| clarkb | greghaynes: https://review.openstack.org/#/c/156356/1/tools/build_wheels.sh doesn't do anything with PIP_WHEEL_DIR | 18:11 |
| *** carl_baldwin has quit IRC | 18:11 | |
| *** emagana has joined #openstack-infra | 18:12 | |
| greghaynes | ah, was looking at https://review.openstack.org/#/c/156356/1/stackrc | 18:12 |
| clarkb | stackrc does it, but that allows you to get out of sync | 18:12 |
| openstackgerrit | Khai Do proposed openstack-infra/project-config: Use job templates for gerritlib builds https://review.openstack.org/156068 | 18:12 |
| clarkb | maybe that line in the script should just be removed? | 18:12 |
| clarkb | ya I think that is what I will suggest | 18:13 |
| *** dhritishikhar_ has quit IRC | 18:13 | |
| *** ihrachyshka has joined #openstack-infra | 18:13 | |
| *** bknudson has quit IRC | 18:14 | |
| *** andreykurilin_ has quit IRC | 18:14 | |
| *** annegent_ has quit IRC | 18:14 | |
| *** carl_baldwin has joined #openstack-infra | 18:14 | |
| *** andreykurilin_ has joined #openstack-infra | 18:14 | |
| hogepodge | fungi where should I get started on the board team and defcore project? ttx uploaded this but has it marked wip https://review.openstack.org/#/c/155738/ | 18:14 |
| *** spzala has joined #openstack-infra | 18:15 | |
| *** emagana has quit IRC | 18:15 | |
| *** mriedem1 is now known as mriedem | 18:15 | |
| fungi | hogepodge: probably there--taking a look now | 18:15 |
| *** HeOS has quit IRC | 18:16 | |
| *** bknudson has joined #openstack-infra | 18:16 | |
| fungi | hogepodge: ahh, yep, so next will be to write a new repo creation change which depends on that governance change, using the same repo name | 18:16 |
| fungi | hogepodge: i can do that real quick unless you want to take a stab at it | 18:16 |
| hogepodge | I like real quick. I'd probably flail. | 18:17 |
| hogepodge | fungi: ^^ | 18:17 |
| fungi | i just flail fast | 18:17 |
| fungi | ;) | 18:17 |
| hogepodge | :-D | 18:17 |
| AJaeger | jeblair, clarkb, fungi: https://review.openstack.org/#/c/156068/ has a depends-on - but that change-id exists twice, once for the job that is needed and once for an abandoned job. | 18:17 |
| AJaeger | Will this work? | 18:17 |
| clarkb | AJaeger: no you cannot depend on an abandoned change iirc | 18:18 |
| clarkb | AJaeger: well you can but you won't merge | 18:18 |
| *** achanda has joined #openstack-infra | 18:18 | |
| AJaeger | clarkb: thanks for confirming my guess | 18:18 |
| *** thedodd has quit IRC | 18:18 | |
| *** ghostpl_ has joined #openstack-infra | 18:21 | |
| *** pfalleno1 has quit IRC | 18:21 | |
| *** jp_at_hp has quit IRC | 18:22 | |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/project-config: Add job runtime system package and database setup https://review.openstack.org/156698 | 18:22 |
| *** pfallenop has joined #openstack-infra | 18:22 | |
| fungi | hogepodge: that's ^ not it by the way | 18:22 |
| fungi | just switching gears ;) | 18:22 |
| clarkb | fungi: ooh is that working happily now? | 18:23 |
| *** vigneshvar has joined #openstack-infra | 18:23 | |
| fungi | clarkb: no, wip. i need to expand the packages list | 18:23 |
| clarkb | kk | 18:23 |
| zaro | AJaeger, clarkb : wondering if creating a new change id for the *good* change is a viable workaround? | 18:24 |
| fungi | clarkb: though the db prep and creation macros should be working now | 18:24 |
| fungi | clarkb: tested that the commands do what i think they should anyway, getting closer to trying nova unit tests to see what's still lacking | 18:24 |
| *** changbl has joined #openstack-infra | 18:24 | |
| openstackgerrit | Cedric Brandily proposed openstack-infra/git-review: Support authentication in run_http_exc https://review.openstack.org/124410 | 18:24 |
| AJaeger | zaro: that means abandon the good change and resubmit of it. Yes, one option | 18:24 |
| clarkb | zaro: I think that is the only workaround | 18:24 |
| jeblair | zaro, AJaeger: this is an interesting edge-case for CRD | 18:24 |
| *** sigmavirus24 is now known as sigmavirus24_awa | 18:25 | |
| clarkb | fungi: nice | 18:25 |
| jeblair | zaro, AJaeger: perhaps the desired behavior is actually just to ignore abandoned CRD. | 18:25 |
| fungi | clarkb: i _think_ i'd like to look into resurrecting lifeless's bindep project to replace most/all of the base-packages macro | 18:25 |
| jeblair | i think it's a question with no clear right answer .. | 18:25 |
| clarkb | jeblair: ya I think either way you end up with confusion and potentially extra work | 18:25 |
| clarkb | jeblair: it would be surprising if say a change merged without a dependency because that dep was abandoned | 18:26 |
| fungi | clarkb: since that way we could stash info on what system packages are needed for a to run tests inside the individual project repos | 18:26 |
| clarkb | fungi: since they can/will differ. that makes sense to me | 18:26 |
| jeblair | clarkb: yep, otoh, in this case, if you accidentally create a change with the id and realize it's no longer needed, you have to make a new change for any changes that _should_ land | 18:26 |
| fungi | clarkb: and also prevents us from installing nova system requirements to run keystone tests, et cetera | 18:27 |
| fungi | so some potential streamlining to recoup the efficiency loss from job run-time package installation | 18:27 |
| fungi | pre-caching these to speed up installs is also going to be a separate exercise | 18:28 |
| jeblair | fungi: bindep is in-tree specification of binary deps? | 18:28 |
| fungi | jeblair: yeah, with some cross-distro support | 18:28 |
| *** jlibosva has quit IRC | 18:28 | |
| jeblair | fungi: with something like that, the node prep scripts can crawl all gerrit repos and pre-cache the superset | 18:29 |
| fungi | agreed | 18:29 |
| fungi | it's tractible, just need to flesh out the details | 18:29 |
| clarkb | or maybe we have a list of global system bindep deps in requirements repo? | 18:29 |
| fungi | http://git.openstack.org/cgit/stackforge/bindep is already in gerrit btw | 18:29 |
| clarkb | I don't like either option a whole lot but both should work | 18:29 |
| fungi | not sure how working it is yet | 18:29 |
| jeblair | clarkb: i like in-tree better than global | 18:30 |
| zaro | jeblair: what if we can depend on multiple things? like depend on change id & topic? | 18:30 |
| jeblair | clarkb: i don't see a reason we need a single review bottleneck on that | 18:30 |
| fungi | yeah, global solves pip/pypi problems that distro packaging won't have | 18:30 |
| fungi | distros already solve that problem for themselves | 18:30 |
| fungi | unless we start mixing and matching external package repositories or something | 18:30 |
| *** vigneshvar has quit IRC | 18:31 | |
| fungi | which, you know, we should just try to avoid doing in gratuitous amounts | 18:31 |
| clarkb | jeblair: well possibly because fedora vs ubuntu vs centos don't support a required C package | 18:31 |
| *** sigmavirus24_awa is now known as sigmavirus24 | 18:31 | |
| clarkb | jeblair: but in theory the python dep review would catch that | 18:31 |
| *** carl_baldwin has quit IRC | 18:32 | |
| fungi | also keeping the bindep manifests in the projects means that changes to it self-test with the jobs using those | 18:33 |
| clarkb | pc_m: your functional job should be voting soon :) | 18:33 |
| *** koolhead17 has joined #openstack-infra | 18:34 | |
| *** cebruns_ is now known as cebruns | 18:34 | |
| clarkb | fyi I have been avoiding merging any new project changes while the ansible bug is outstanding | 18:35 |
| pc_m | clarkb: Hi. I was holding that review as WIP, because of some other concerns. I sent email to ML asking whether we should change directory layout and have a second functional test job to handle StrongSwan implementation. | 18:35 |
| openstackgerrit | Khai Do proposed openstack-infra/project-config: Use job templates for gerritlib builds https://review.openstack.org/156068 | 18:35 |
| pc_m | clarkb: Please look at the thread...http://lists.openstack.org/pipermail/openstack-dev/2015-February/057138.html Would love to get some feedback on how we should proceed on this. | 18:36 |
| fungi | hogepodge: do you have any initial git content you need imported into openstack/defcore so we can preserve its commit history, or is this green-field? | 18:36 |
| clarkb | pc_m: that change was no WIP in gerrit... | 18:37 |
| pc_m | oh | 18:37 |
| pc_m | I might have forgot to mark it so. | 18:37 |
| clarkb | pc_m: I have removed my approval please mark it WIP if that is what you want | 18:38 |
| openstackgerrit | yolanda.robla proposed openstack-infra/nodepool: WIP: Not even close to finished - use shade https://review.openstack.org/156322 | 18:38 |
| pc_m | clarkb: Thanks. I just did. | 18:38 |
| pc_m | clarkb: Please look at the thread on ML. Maybe this can proceed, but I'd like to hold to see if things are OK as-is. | 18:39 |
| *** devlaps has joined #openstack-infra | 18:40 | |
| *** martingc_ is now known as martingc | 18:40 | |
| clarkb | pc_m: assuming the current driver system works in that test I don't see why you shouldn't test it | 18:40 |
| clarkb | pc_m: then your transition to the other driver can be self testing (or as much as it can be) | 18:40 |
| pc_m | clarkb: Current driver does pass. Will need new job for new driver (as different env setup). | 18:42 |
| pc_m | clarkb: Not sure if directory structure can be reused as-is or if it should be organized differently. | 18:42 |
| pc_m | clarkb: question 8 in my email. | 18:43 |
| *** vigneshvar has joined #openstack-infra | 18:43 | |
| clarkb | pc_m: I don't know how to ansewr that, I think it is up to the project to decide how tests are organized | 18:44 |
| reed | clarkb, I added "Many IRC channels are logged and recordings are publicly accessible. If you're concerned about privacy consider using a cloak, tor, hide your real name and be mindful not to write sensitive data in these channels. " to https://wiki.openstack.org/wiki/IRC | 18:44 |
| openstackgerrit | K Jonathan Harker proposed openstack-infra/zuul: Add a remote url override location https://review.openstack.org/76057 | 18:44 |
| *** sputnik13 has joined #openstack-infra | 18:44 | |
| reed | with links to freenode FAQ | 18:45 |
| clarkb | reed: sounds good | 18:45 |
| reed | clarkb, not sure we need to put a notice in /topic, too... as you said, IRC should always be considered public anyway | 18:46 |
| openstackgerrit | Merged openstack-infra/project-config: We have a stable/juno branch ready https://review.openstack.org/156262 | 18:49 |
| reed | mrmartin, is this https://storyboard.openstack.org/#!/story/1082784 the story you need to add to https://review.openstack.org/#/c/154061/? | 18:49 |
| *** devlaps has quit IRC | 18:49 | |
| anteaya | so neutron-vpnaas is using this line to pull in neutron as a dependency for testing: http://git.openstack.org/cgit/openstack/neutron-vpnaas/tree/tox.ini#n13 | 18:50 |
| anteaya | how often does an egg get updated? | 18:50 |
| hogepodge | fungi we can call this is green-field. There is existing code in refstack, but we we'll want to start a new repo | 18:50 |
| clarkb | anteaya: in that case since its via the git repo it is always as up to date as the source repo | 18:50 |
| anteaya | great | 18:50 |
| anteaya | thank you | 18:50 |
| fungi | hogepodge: we can use git filter-branch to split out any files from refstack on which you want to maintain commit history | 18:51 |
| openstackgerrit | Merged openstack-infra/project-config: Add the docs-on-rtfd job to stackforge/vmtp https://review.openstack.org/156341 | 18:51 |
| hogepodge | fungi ooh, yeah, that would be nice. | 18:52 |
| sdague | clarkb: any chance you could take a look at gerrit memory bloat? it feels like it's started to get slow again | 18:53 |
| fungi | hogepodge: can you get me a list of files or subtrees you want split out? | 18:53 |
| hogepodge | fungi everything in this directory https://github.com/stackforge/refstack/tree/master/defcore | 18:53 |
| mrmartin | reed ah, we have two stories for that: https://storyboard.openstack.org/#!/story/2000158 | 18:53 |
| jeblair | sdague: we identified memory bloat as the cause? | 18:53 |
| fungi | hogepodge: perfect--can do | 18:53 |
| hogepodge | \o/ | 18:53 |
| sdague | jeblair: I believe that's what clarkb said the last time | 18:53 |
| reed | mrmartin, interesting, I get: 401: GET /api/v1/stories/2000158: No error details available. | 18:54 |
| reed | oh, now it works | 18:54 |
| mrmartin | awesome, when I made this story, storyboard was down | 18:54 |
| clarkb | jeblair: sdague: gerrit was bumping up against its 8GB jvm heap size last time | 18:54 |
| krotscheck | eh? | 18:54 |
| *** carl_baldwin has joined #openstack-infra | 18:54 | |
| reed | mrmartin, https://storyboard.openstack.org/#!/story/1082784 is listed as "merged" anyway | 18:55 |
| clarkb | let me grab a link to the java melody graphs | 18:55 |
| *** tqtran has quit IRC | 18:55 | |
| fungi | hogepodge: is there any stratification of maintainers for the repo (just a core group)? do you plan on tagging repo state at various points (maybe via a separate release group)? | 18:55 |
| * reed dances to the java melody | 18:55 | |
| clarkb | jeblair: https://review.openstack.org/monitoring?part=graph&graph=usedMemory&period=mois | 18:55 |
| jhesketh | Morning | 18:55 |
| mrmartin | reed, it was a different task: Create a git repo in gerrit in which customization plugin and skin can go. << just for the askbot-theme gerrit and project | 18:55 |
| reed | mrmartin, ACK | 18:56 |
| clarkb | jhesketh: good morning | 18:56 |
| pc_m | clarkb: Yeah, looking (hoping) for some guidance on what may be a good way to organize the test dir to support two functional jobs | 18:56 |
| hogepodge | fungi there could be. board members should have more power over core members, especially in appointing core members. core members should be able to merge though. | 18:56 |
| anteaya | jhesketh: morning | 18:57 |
| hogepodge | fungi we can handle that socially, though, if it adds too much complexity | 18:57 |
| hogepodge | fungi know the meeting is soon, so we can continue later. | 18:57 |
| krotscheck | mrmartin, reed: That issue you two encountered _might_ be fixed by this patch, it’s still pending: https://review.openstack.org/#/c/155943/ | 18:57 |
| fungi | hogepodge: okay, we may have to noodle on how the delegation works, but it won't impact the patch i'm writing in that case | 18:57 |
| *** mriedem has quit IRC | 18:58 | |
| openstackgerrit | Khai Do proposed openstack-infra/jenkins-job-builder: Change test folder name to 'fixture' for consistency https://review.openstack.org/152664 | 18:59 |
| *** rcarrillocruz has quit IRC | 18:59 | |
| clarkb | jhesketh: fwiw I haven't had a chance to look into the upload slowness yet | 18:59 |
| clarkb | jhesketh: hopefully this afternoon I will take a look | 19:00 |
| clarkb | jhesketh: I was wondering if requests is making a single POST with all of those files sequentially rather than 10x POSTs | 19:00 |
| fungi | hogepodge: is anyone going to be committing updates to that directory in refstack in the next few days? we may want to consider that content effectively frozen while the new repo is created | 19:00 |
| jeblair | meeting time! | 19:00 |
| cebruns | o/ | 19:01 |
| jeblair | cebruns: we're in #openstack-meeting | 19:01 |
| jhesketh | clarkb: all good. Yeah I'm not sure and I didn't get a chance either :-( | 19:01 |
| cebruns | jeblair: thanks - wrong tab. :) | 19:01 |
| *** tiswanso has quit IRC | 19:01 | |
| *** mriedem has joined #openstack-infra | 19:01 | |
| *** tiswanso_ has joined #openstack-infra | 19:02 | |
| *** rcarrillocruz has joined #openstack-infra | 19:03 | |
| *** rverchikov has quit IRC | 19:03 | |
| *** sputnik13 has quit IRC | 19:04 | |
| hogepodge | fungi yes, we want this to land, but it can wait https://review.openstack.org/#/c/155866/ | 19:04 |
| *** Rockyg has joined #openstack-infra | 19:04 | |
| *** rverchikov has joined #openstack-infra | 19:04 | |
| *** sputnik13 has joined #openstack-infra | 19:04 | |
| *** vigneshvar has quit IRC | 19:04 | |
| *** ihrachyshka has quit IRC | 19:04 | |
| fungi | hogepodge: any chance you want to rush it in before i make the split-out copy? | 19:04 |
| hogepodge | fungi there's no rushing with this project | 19:05 |
| fungi | hah | 19:05 |
| fungi | okay, in that case once we get the new repo created, that change can be abandoned and reproposed to it | 19:06 |
| *** pfallenop has quit IRC | 19:09 | |
| *** pfallenop has joined #openstack-infra | 19:10 | |
| openstackgerrit | Merged openstack-infra/project-config: Remove jenkins console-log publisher from infra https://review.openstack.org/156520 | 19:11 |
| *** ssam2 has quit IRC | 19:13 | |
| *** pelix has quit IRC | 19:14 | |
| *** annegent_ has joined #openstack-infra | 19:15 | |
| jogo | marun: ping, any updates for how to get around the non-voting to voting issue for skip-if? | 19:15 |
| marun | jogo: I have a wip patch I'm about to post for review by jeblair. | 19:15 |
| marun | jogo: so, in process :) | 19:15 |
| jogo | marun: cool | 19:16 |
| jogo | marun: also can you check https://review.openstack.org/#/c/156214/1/zuul/layout.yaml,cm to make sure it doesn't hit that bug? | 19:16 |
| jogo | I think it may | 19:16 |
| jogo | sdague: ^ | 19:17 |
| *** abhirc has quit IRC | 19:17 | |
| AJaeger | jogo: I agree, it would hit jobs ending with "-nv" | 19:18 |
| sdague | for someone that wasn't around for the bug, what is the core issue? | 19:18 |
| AJaeger | jogo: you might be lucky that none match today but no promises for the future | 19:19 |
| fungi | marun: jogo: once the fix is merged to the zuul repo, the layout job run on your project-config change should show whether it sets any nonvoting jobs to voting | 19:19 |
| fungi | it's just an informational job though, not gating | 19:19 |
| *** openstackgerrit has quit IRC | 19:20 | |
| *** openstackgerrit has joined #openstack-infra | 19:20 | |
| *** vigneshvar has joined #openstack-infra | 19:20 | |
| *** annegent_ has quit IRC | 19:20 | |
| *** ghostpl_ has quit IRC | 19:20 | |
| jogo | fungi: ahh, would it 'fail' iff it changes anything back to voting? or do we have to go log spelunking | 19:21 |
| jogo | clarkb: FYI https://review.openstack.org/#/c/156434/2/files/debs/nova,cm | 19:21 |
| jogo | clarkb: can I remove anything else from that? | 19:21 |
| clarkb | jogo: where is libvirt being installed now that you remove it from that list? | 19:21 |
| *** marun_ has joined #openstack-infra | 19:21 | |
| *** abhirc has joined #openstack-infra | 19:21 | |
| jogo | clarkb: its pip installable now | 19:22 |
| clarkb | jogo: but not in an requirements lists | 19:22 |
| clarkb | jogo: the python-mysql* packages and whatever postgres packages are used would also be candidates for removal | 19:22 |
| fungi | jogo: no, you'd just need to look through the parsed output of the layout in the console log to see if there are any -nv jobs without voting:false set | 19:22 |
| jogo | clarkb: huh, that is a great question | 19:23 |
| *** marun has quit IRC | 19:23 | |
| openstackgerrit | Maru Newby proposed openstack-infra/zuul: WIP Stop metajob application of undefined attrs https://review.openstack.org/156716 | 19:23 |
| marun_ | jeblair: ^^ | 19:23 |
| *** markmcclain has quit IRC | 19:24 | |
| *** markmcclain has joined #openstack-infra | 19:24 | |
| jogo | clarkb: oh libvirt is still apt-get installed http://logs.openstack.org/34/156434/2/check/check-tempest-dsvm-full/6b69884/logs/devstacklog.txt.gz#_2015-02-17_03_41_15_298 | 19:25 |
| openstackgerrit | Ramy Asselin proposed openstack-infra/puppet-zuul: Lockfile removed from requirements https://review.openstack.org/156721 | 19:30 |
| sdague | jogo: yes, nothing changed about that | 19:30 |
| dansmith | I know this is silly, but can I submit a config change to adjust the ordering of the jobs in the oslo.versionedobjects tests? | 19:30 |
| sdague | jogo: so... I was doing a more comprehensive purge today - https://review.openstack.org/#/c/156594/ | 19:31 |
| sdague | I didn't notice your thing existed | 19:31 |
| *** bhuvan_ has quit IRC | 19:31 | |
| fungi | dansmith: can you elaborate? you mean the order in which the results are posted in a gerrit comment? | 19:32 |
| jogo | sdague: happy to abandon mine | 19:32 |
| dansmith | fungi: yeah | 19:32 |
| jogo | sdague: you can get rid of a few more things though | 19:33 |
| fungi | dansmith: i'd have to look at whether zuul uses an ordered data type for those or not | 19:33 |
| jogo | sdague: see https://review.openstack.org/#/c/156434/2/files/debs/nova,cm | 19:33 |
| sdague | jogo: well, getting rid of the stuff with native libs really negatively impacts one of the primary use cases, which is using devstack for reals outside of the gate | 19:34 |
| dansmith | fungi: they always seem very consisent | 19:34 |
| dansmith | fungi: I'll check to see if the layout matches the order they're reported in | 19:34 |
| jogo | sdague: because it will take a few minutes longer to run? | 19:34 |
| clarkb | dansmith: fungi its determined by the order ofthe jobs list in zuul's layout.yaml | 19:34 |
| fungi | clarkb: in that case the job-templates complicate that pretty heavily i suspect | 19:34 |
| dansmith | clarkb: yeah, seemed to be very consistent | 19:34 |
| openstackgerrit | Clint 'SpamapS' Byrum proposed openstack-infra/reviewstats: Deactivate Robert Collins from TripleO https://review.openstack.org/156724 | 19:35 |
| sdague | jogo: and do all the dev package installs | 19:35 |
| jogo | sdague: mixing system and pip packages is an even worse | 19:35 |
| clarkb | fungi: yup | 19:35 |
| *** Rockyg has quit IRC | 19:35 | |
| dansmith | clarkb: is it too insane for me to submit a reordering change? | 19:36 |
| sdague | jogo: can we do this in phases? because I'd rather do all the slam dunks here, see what the fallout might be, then come up with a binaries story later | 19:36 |
| AJaeger | dansmith: seems to be a virus this week ;) https://review.openstack.org/#/c/154291/ | 19:36 |
| fungi | dansmith: so based on it being ordered based on parsing the yaml combined with expanding project-templates, it will likely be fiddly | 19:36 |
| *** Rockyg has joined #openstack-infra | 19:36 | |
| jogo | sdague: yes we can do it in phases. that is why I didn't -1 your patch | 19:36 |
| AJaeger | dansmith: that was last week ;) | 19:37 |
| dansmith | fungi: I'm fine with the ordering I can achieve even with the templates | 19:37 |
| dansmith | fungi: by moving a single line in my config :) | 19:37 |
| dansmith | AJaeger: OCD is not a virus :P | 19:37 |
| sdague | jogo: and I'd like to wait until dtroyer's wheel cache bits are in place before we touch these | 19:37 |
| fungi | dansmith: sure, if it's important to you, put in a patch. i'd have spent less time to review than to discuss ;) | 19:37 |
| sdague | these == stuff we have to compile | 19:37 |
| AJaeger | dansmith: OCD? | 19:37 |
| jogo | sdague: kk | 19:37 |
| dansmith | fungi: sorry :/ | 19:38 |
| openstackgerrit | Dan Smith proposed openstack-infra/project-config: Reorder python jobs in oslo.versionedobjects https://review.openstack.org/156727 | 19:38 |
| dansmith | AJaeger: obsessive compulsive disorder | 19:38 |
| marun_ | jogo, sdague: are we removing distro-supplied python packages? | 19:38 |
| dansmith | AJaeger: being overly concerned with minor details | 19:38 |
| jogo | marun_: IMHO we should | 19:38 |
| dansmith | fungi: https://review.openstack.org/156727 | 19:38 |
| dansmith | fungi: I'll sleep much better with that :D | 19:38 |
| *** mriedem has quit IRC | 19:39 | |
| marun_ | jogo, sdague: I'm filtering them out for the neutron functional job because we don't use them, but that means there's a difference between the dependencies of functional and integration neutron | 19:39 |
| *** marun_ is now known as marun | 19:39 | |
| marun | I'm more than happy to see us use just python packages | 19:39 |
| marun | and let the distros worry about validating their packages | 19:39 |
| sdague | marun: yeh, so I have a patch up to purge everything that's 100% python | 19:39 |
| AJaeger | dansmith: spent the team a round of beer in vancouver ;) and don't expect quick reviews | 19:40 |
| sdague | marun: https://review.openstack.org/#/c/156594/ | 19:40 |
| marun | sdague: /url 1 | 19:40 |
| *** annegent_ has joined #openstack-infra | 19:40 | |
| marun | oops | 19:40 |
| marun | sdague: nice | 19:40 |
| marun | but only for deb? | 19:40 |
| marun | are you intending to do the same for rpm distros? | 19:40 |
| dansmith | AJaeger: I owe more than I can afford already :/ | 19:40 |
| sdague | marun: well mostly I wanted to see the test fallout | 19:40 |
| jogo | marun: we do gate on RPM stuff so in theory we can do that too | 19:40 |
| sdague | yes, the intent is to do it on the rpm side as well | 19:40 |
| *** mriedem has joined #openstack-infra | 19:41 | |
| marun | sdague: What's the rational for not stripping all python deps, not just the pure ones? | 19:41 |
| sdague | marun: because stripping the others are more complicated, because that means you need to make sure you install all the packages to make them compile correctly all the time | 19:41 |
| marun | sdague: I'm working on a solution for that. | 19:42 |
| sdague | and I haven't done all the -dev package analysis there | 19:42 |
| marun | sdague: but did I hear that dtroyer is as well? | 19:42 |
| sdague | marun: https://review.openstack.org/#/c/156356/ | 19:42 |
| AJaeger | dansmith: we spent so much time discussing that I just approved it ;) | 19:42 |
| marun | sdague: I have a prototype that will take a requirements list and a mapping of requirements to distro deps, and try to install each requirement in a bare distro + known requirements | 19:43 |
| marun | sdague: it uses docker for the base distro | 19:43 |
| dansmith | AJaeger: see, that's the strategy.. discuss so much that approving the patch is the simplest way to stop the bleeding :) | 19:43 |
| dansmith | AJaeger: and thanks :) | 19:43 |
| AJaeger | dansmith: next time I hide better ;) | 19:43 |
| *** cdent has quit IRC | 19:43 | |
| dansmith | heh | 19:43 |
| marun | sdague: I went through neutron's requirements and test requirements, could go through the global requirements and do the same thing easily enough | 19:43 |
| AJaeger | dansmith: glad that we could make you happy with such a trivial patch ;) | 19:43 |
| fungi | marun: i'm also planning to see if http://git.openstack.org/cgit/stackforge/bindep will work for getting lists of non-python deps tracked in individual repos | 19:44 |
| marun | dtroyer: hi | 19:44 |
| dansmith | AJaeger: ...this time :) | 19:44 |
| sdague | marun: so I think in this case it's not clear that we want to use the distro bits | 19:44 |
| fungi | marun: as part of our current effort to stop having different worker types for devstack and unit test jobs | 19:44 |
| openstackgerrit | Ian Wienand proposed openstack-infra/system-config: Use nodepool 'generate-log-config' command to generate log config https://review.openstack.org/156457 | 19:44 |
| openstackgerrit | Doug Hellmann proposed openstack-infra/release-tools: Rename oslo_release.sh to library_release.sh https://review.openstack.org/151285 | 19:44 |
| marun | fungi: bindep is only for ancillary deps right now, though | 19:44 |
| openstackgerrit | Merged openstack/requirements: Bump pycadf to 0.8.0 https://review.openstack.org/155652 | 19:45 |
| marun | fungi: I'm talking about being able to map python dep -> distro requirements & other python deps | 19:45 |
| *** Rockyg_ has joined #openstack-infra | 19:45 | |
| *** pfallenop has quit IRC | 19:46 | |
| marun | fungi: it's simple enough, I need to post it on github this week | 19:46 |
| openstackgerrit | Merged openstack/requirements: update python-heatclient>=0.3.0 https://review.openstack.org/154989 | 19:46 |
| *** pfallenop has joined #openstack-infra | 19:46 | |
| dtroyer | marun: hey | 19:47 |
| clarkb | fungi: marun's thing may more completely solve your issue | 19:47 |
| *** gulic has quit IRC | 19:47 | |
| clarkb | fungi: if it can just feed off of the requirements.txt and test-requirements.txt directly | 19:47 |
| *** Rockyg_ has quit IRC | 19:47 | |
| *** Rockyg has quit IRC | 19:47 | |
| marun | dtroyer: I hashed together something that maps python deps to distro deps and validates them on clean hosts using docker | 19:47 |
| *** Rockyg has joined #openstack-infra | 19:48 | |
| marun | dtroyer: I was wondering if it would be useful in the work you're doing here: https://review.openstack.org/#/c/156356/ | 19:48 |
| marun | dtroyer: to make the dependency mapping more explicit | 19:48 |
| openstackgerrit | Clark Boylan proposed openstack-infra/nodepool: Remove unnecessary path joining for dib filenames https://review.openstack.org/139704 | 19:48 |
| *** otter768 has joined #openstack-infra | 19:48 | |
| fungi | marun: yeah, tangential to what i'm doing, just want to make sure if you have better suggestions... and also trying not to step on your toes | 19:48 |
| marun | fungi: No toes to step on yet, though I have similar goals to you. | 19:49 |
| greghaynes | mordred: clarkb idea about the ssh key issue for public images - why not just make two images? Just run a post-build operation that copies the built image and adds a joejob password to the user on the image | 19:49 |
| dtroyer | marun: the wheel cache bits pre-build PyPI packages that we don't want to have to build 5 times to create 5 venvs. I'm really not considering distro python packages at all. | 19:49 |
| marun | fungi: I'd definitely like to see individual repos be responsible for currating their dependencies. | 19:49 |
| marun | dtroyer: not distro python packages | 19:49 |
| marun | dtroyer: distro non-python packages | 19:50 |
| fungi | marun: also my plan makes them self-testing | 19:50 |
| marun | dtroyer: you're already doing that mapping, but it's very loose | 19:50 |
| dtroyer | marun: ah, the pre-reqs needed to build them? | 19:50 |
| fungi | which means less chance of breakage from infra changing package lists in our repos | 19:50 |
| marun | dtroyer: e.g. https://review.openstack.org/#/c/156356/1/files/venv-requirements.txt | 19:50 |
| sdague | marun: so ... are you basically building an openstack package manager? :) | 19:50 |
| marun | dtroyer: yes | 19:50 |
| *** mriedem has quit IRC | 19:50 | |
| mordred | greghaynes: that's entirely too sane | 19:50 |
| sdague | which, I actually support as a concept, because we seem to keep dancing around it and putting that logic in other places | 19:51 |
| openstackgerrit | patrick-crews proposed openstack-infra/project-config: Increase py26 timeouts for nova-stable https://review.openstack.org/156734 | 19:51 |
| * greghaynes gets out of here with these reasonable ideas | 19:51 | |
| marun | dtroyer: so there would be an explicit mapping between, say, mysql-python and the underlying distros build deps | 19:51 |
| dtroyer | marun: right, and that is loose, and doesn't handle things like multi-node DevStack and compute-only nodes, etc. | 19:51 |
| dtroyer | I'm just focused on getting services into venvs right now | 19:51 |
| marun | dtroyer: ok. no overlap as of yet, then. | 19:51 |
| marun | sdague: maybe? | 19:51 |
| * greghaynes -> train | 19:51 | |
| marun | sdague: I'd just like the relationships between our projects, python deps and distro deps to be more explicit so we can manage it better. | 19:52 |
| marun | sdague: I'm kind of surprised that python doesn't already have a mechanism for doing this. | 19:52 |
| marun | sdague: It seems obvious | 19:52 |
| *** mriedem has joined #openstack-infra | 19:52 | |
| marun | fungi: will ping you later, I'm interested to hear what you're planning especially the self testing. | 19:53 |
| sdague | marun: so from what I can tell the python answer is wheels | 19:53 |
| *** otter768 has quit IRC | 19:53 | |
| sdague | except wheels that work generically on linux are hard | 19:53 |
| sdague | because of lib drift, in a way that mac/windows it is not | 19:53 |
| *** mrunge has quit IRC | 19:54 | |
| clarkb | sdague: not to mention abi changes. see that f22 is making the big leap for c++11 support in g++? | 19:54 |
| sdague | clarkb: right | 19:54 |
| *** devanand_ has quit IRC | 19:55 | |
| marun | sdague: wheels is the result | 19:56 |
| marun | sdague: Hmmmm | 19:57 |
| sdague | marun: right, but for mac/windows because they have such a limitted toolchain, wheels is the starting point | 19:57 |
| sdague | binaries just run | 19:57 |
| marun | sdague: I seem to recall eggs promising the same thing | 19:57 |
| marun | sdague: I used to use them for that very reason | 19:57 |
| marun | sdague: So maybe I'm simply trying to solve a higher-order problem that openstack will shortly not have. | 19:58 |
| marun | sdague: but who distributes the wheels? | 19:58 |
| marun | sdague: are they autobuilt and distributed on pypi today and I don't know about it? | 19:58 |
| *** sdake has quit IRC | 19:58 | |
| marun | ack, bbiab | 19:59 |
| sdague | marun: they can be - https://pypi.python.org/pypi/numpy/1.9.1 | 19:59 |
| sdague | although it looks like it's up to the author to push them, so you only get one platform - https://pypi.python.org/pypi/lxml/3.4.2 | 20:01 |
| *** sabeen3 has joined #openstack-infra | 20:02 | |
| clarkb | ya pypi isn't building them in a farm or anything | 20:02 |
| clarkb | its a thing you can do when uploading (we always upload an sdist and wheel wherever possible) | 20:02 |
| sdague | yeh, though in the noarch case it's less interesting | 20:03 |
| clarkb | yup | 20:03 |
| *** mbacchi has quit IRC | 20:03 | |
| sdague | yeh, so without a build farm, it doesn't really solve much | 20:03 |
| *** sabeen has quit IRC | 20:03 | |
| *** mbacchi has joined #openstack-infra | 20:04 | |
| *** sdake has joined #openstack-infra | 20:05 | |
| openstackgerrit | Merged openstack-infra/project-config: Reorder python jobs in oslo.versionedobjects https://review.openstack.org/156727 | 20:05 |
| *** bhuvan has joined #openstack-infra | 20:07 | |
| *** dcramer_ has joined #openstack-infra | 20:07 | |
| fungi | sdague: though you can upload multiple wheels of the same release version targeting different platforms | 20:07 |
| clarkb | fungi: but I don't think it knows how to distinguish among linux distros | 20:08 |
| asselin | jeblair, just to reconfirm, third-party is now a priority effort? That refers to this, right? https://review.openstack.org/#/c/139745/ | 20:08 |
| clarkb | fungi: so you have linux, OSX, windows, maybe *BSD | 20:08 |
| fungi | clarkb: right, it's mostly for py2.x vs 3.x pure python wheels, or osx vs windows | 20:08 |
| jeblair | asselin: yes, and we'll probably combine the openstack_project refactor in there too (it's partially related) | 20:09 |
| sdague | yeh, the numpy and lxml links are interesting data points | 20:09 |
| asselin | jeblair, great! thanks! | 20:09 |
| sdague | the naming, leaves a lot to be desired | 20:09 |
| mordred | clarkb, sdague: basically, uploading binary wheels for linux to pypi is the worst idea ever | 20:09 |
| openstackgerrit | Dmitry Tantsur proposed openstack/requirements: Do not break on projects without setup.cfg https://review.openstack.org/156220 | 20:09 |
| *** sdake has quit IRC | 20:09 | |
| mordred | and I believe that pip will not use those by default current | 20:09 |
| mordred | you have to tell it "no, seriously, screw my entire computer" | 20:09 |
| sdague | numpy-1.9.1-cp34-cp34m-macosx_10_6_intel.macosx_10_9_intel.macosx_10_9_x86_64.macosx_10_10_intel.macosx_10_10_x86_64.whl (md5) | 20:10 |
| fungi | gotta love that filename | 20:10 |
| clarkb | really what wheels should've been were like windows installers where every dep is vendored | 20:10 |
| *** vigneshvar has quit IRC | 20:10 | |
| fungi | clarkb: that also sounds like most mac os/os x applications | 20:10 |
| *** sabeen3 has quit IRC | 20:11 | |
| fungi | also sco unix, rpath linux... | 20:11 |
| clarkb | we could use nix for everything | 20:11 |
| *** sabeen1 has joined #openstack-infra | 20:11 | |
| fungi | right, that recent nixos thing | 20:11 |
| fungi | i was trying to remember the name of it | 20:11 |
| *** mbacchi has quit IRC | 20:13 | |
| *** tiswanso_ has quit IRC | 20:13 | |
| clarkb | you can just use nix the package manager and get all the fun on whatever distro too | 20:13 |
| *** mbacchi has joined #openstack-infra | 20:14 | |
| dcramer_ | Anybody have experience with Ansible's OpenStack inventory script? I'm unclear on how to make it aware of my credentials. https://github.com/ansible/ansible/blob/devel/plugins/inventory/openstack.py | 20:14 |
| nibalizer | sdague: impressive | 20:14 |
| * jeblair -> lunch | 20:14 | |
| openstackgerrit | Pranav Salunke proposed openstack-infra/project-config: Updates training guides jobs https://review.openstack.org/152118 | 20:15 |
| AJaeger | I'd appreciate an approval for this patch to infra-manual: https://review.openstack.org/154065 | 20:15 |
| nibalizer | clarkb: or whomever so i have a question: when a change to e.g. nova gets pushed up, it runs devstack, does it do that on ubuntu and centos, or just one? | 20:15 |
| nibalizer | it seems that there are dvsmcentos7 nodes in the nodepool image list | 20:15 |
| fungi | nibalizer: ubuntu trusty for master and juno, precise for icehouse | 20:15 |
| nibalizer | no centos? | 20:15 |
| fungi | nibalizer: no, those are a work in progress i believe. there may be an experimental job trying to use them | 20:16 |
| openstackgerrit | Andreas Jaeger proposed openstack-infra/infra-manual: For openstack projects, add dependency on governance change https://review.openstack.org/156745 | 20:16 |
| *** rfolco_ has quit IRC | 20:16 | |
| AJaeger | here comes by action item from the meeting, anteaya please review ^ | 20:16 |
| clarkb | ya centos7 isn't working right now aiui | 20:17 |
| *** annegent_ has quit IRC | 20:18 | |
| nibalizer | okay, with the stackforge/puppet-* stuff we've got some minimal integration testing running on ubuntu, but there is great desire to test it on centos as well | 20:19 |
| *** dcramer_ has left #openstack-infra | 20:19 | |
| AJaeger | thanks, fungi! | 20:20 |
| openstackgerrit | Merged openstack-infra/infra-manual: Update to reflect recent governance changes https://review.openstack.org/154065 | 20:20 |
| *** ghostpl_ has joined #openstack-infra | 20:21 | |
| *** mrmartin has quit IRC | 20:21 | |
| lifeless | fungi: -yes- :) | 20:21 |
| fungi | lifeless: while i suspect i have context for your affirmation, i am going to pretend it's just a blanket yes for anything else i might have said too ;) | 20:22 |
| lifeless | fungi: naturally | 20:23 |
| *** ghostpl_ has quit IRC | 20:26 | |
| mordred | hey everybody - if dcramer_ comes back, I wrote that inventory and would be happy to help out | 20:27 |
| openstackgerrit | Andreas Jaeger proposed openstack-infra/infra-manual: Fix two typos https://review.openstack.org/156753 | 20:27 |
| clarkb | greghaynes: mordred because the whole point is that anyone can use the image as is and this is already a mostly solved problem (metadata server/cloud init) | 20:27 |
| *** ociuhandu has quit IRC | 20:27 | |
| sdague | nibalizer: so my advice is to get one platform working the way you want it first | 20:27 |
| clarkb | greghaynes: mordred: while I think asking people to run `diskimage-create` isn't a huge deal it would be cool if things just worked as we build them | 20:27 |
| greghaynes | clarkb: sure, you can tell people they have to cloud-init... have you tried to preseed a vm with cloud-init data recently? | 20:28 |
| *** sdake has joined #openstack-infra | 20:28 | |
| clarkb | greghaynes: mordred: its another reason I am still slightly skeptical of mordreds cloud init rewrite | 20:28 |
| clarkb | greghaynes: yes, it involved isos for config drive | 20:28 |
| mordred | clarkb: cloud-init will not help people booting these ina vm | 20:28 |
| clarkb | greghaynes: its like back in the 90s when we had CDs | 20:28 |
| clarkb | mordred: why not? | 20:28 |
| greghaynes | yea, its a giant PITA. IMO we should tell the users to just edit the images then, itll be a lot easier than telling them to cloud-init | 20:28 |
| *** MarkAtwood has joined #openstack-infra | 20:29 | |
| clarkb | mordred: it can set an ssh authorized key file for them | 20:29 |
| clarkb | mordred: and now we don't have to wrry about it | 20:29 |
| mordred | clarkb: I think you are far overestimating the ease of doign that | 20:29 |
| clarkb | greghaynes: except most users already have cloud init | 20:29 |
| mordred | no they don't | 20:29 |
| *** tmcpeak has joined #openstack-infra | 20:29 | |
| clarkb | greghaynes: eg anyone using hpcloud | 20:29 |
| mordred | not in their local vms | 20:29 |
| *** dizquierdo has joined #openstack-infra | 20:29 | |
| clarkb | mordred: my local VMs cloud init... | 20:29 |
| *** vigneshvar has joined #openstack-infra | 20:29 | |
| *** bhuvan has quit IRC | 20:29 | |
| mordred | clarkb: do not use yourself as a typical exampel user :) | 20:30 |
| clarkb | mordred: well thats how ubuntu's published images work | 20:30 |
| clarkb | mordred: so if you consume their images already you have solved this problem | 20:30 |
| mordred | clarkb: when you boot them in what? | 20:30 |
| clarkb | mordred: kvm | 20:30 |
| mordred | ok. how do you set up the metadata? | 20:30 |
| clarkb | mordred: config drive with iso | 20:30 |
| mordred | ok. how do you make the iso? | 20:30 |
| clarkb | mordred: some tool does it for me | 20:30 |
| tmcpeak | hey all - I'm trying to set up tags for versioning. The doc here: http://docs.openstack.org/infra/manual/creators.html#tagging-a-release suggests that I run "git tag -s -m "descriptive message" $version" followed by "git push gerrit $version", but we're using Gerrit reviews. I can't push directly to gerrit. Does anybody know how I should proceed? | 20:30 |
| fungi | for that matter, rackspace's debian images preinstall cloud-init. i had to hack around several tuning changes it was overwriting | 20:31 |
| mordred | and how to you tell it to expose the config-drive as the correct device? | 20:31 |
| clarkb | mordred: again some tool does it all | 20:31 |
| mordred | awesome | 20:31 |
| clarkb | mordred: its actually pretty handy | 20:31 |
| clarkb | I am looking up the tool I was using | 20:31 |
| clarkb | uvtool | 20:31 |
| lifeless | yup | 20:31 |
| anteaya | AJaeger: thanks, commented | 20:31 |
| greghaynes | anyhow, just editing the image is easy, and then in theory users can just download and boot, not sure why that isnt a win? | 20:31 |
| fungi | tmcpeak: if gerrit is refusing the tag push, then your acl probably either lacks an entry to allow that or your account isn't in the group it's allowing that for | 20:31 |
| lifeless | cloud-init really is everywhere :) | 20:32 |
| nibalizer | mordred: ya I hacked the cloud-init from extra drive thing in a completenly different environment | 20:32 |
| nibalizer | its a real thing | 20:32 |
| fungi | tmcpeak: which project? | 20:32 |
| lifeless | unless you use rackspace | 20:32 |
| tmcpeak | fungi: Bandit | 20:32 |
| fungi | tmcpeak: looking now | 20:32 |
| tmcpeak | fungi: thanks! | 20:32 |
| *** cnesa has joined #openstack-infra | 20:32 | |
| *** cnesa has quit IRC | 20:32 | |
| fungi | tmcpeak: the push signed tags permission for that repo is granted to this group https://review.openstack.org/#/admin/groups/444,members | 20:33 |
| clarkb | mordred: it just feels weird to optimize for an already solved problem | 20:33 |
| mordred | clarkb: well, that's neat - I still categorically and completely disagree with you - but like I've said at least 100 times before I think there is 0 chance we're going to sort that out not in person | 20:33 |
| tmcpeak | fungi: ahh, cool | 20:33 |
| clarkb | mordred: I get the concerns but thenI see its a 360 line change of custom stuff to resolve problem and I rethink whether or not its worth it | 20:33 |
| mordred | clarkb: yes. I understand your concerns | 20:34 |
| AJaeger | anteaya: thanks, rewritten | 20:34 |
| fungi | tmcpeak: you should get the current member of that group to modify the membership to include anyone who is expected to be pushing tags (usually kept to an intentionally small subset of your core reviewers) | 20:34 |
| *** r1chardj0n3s_afk is now known as r1chardj0n3s | 20:34 | |
| clarkb | the 360 lines of custom stuff probably works fine, but everyone else will now have to have 2 systems to use our images and their images | 20:34 |
| openstackgerrit | Andreas Jaeger proposed openstack-infra/infra-manual: For openstack projects, add dependency on governance change https://review.openstack.org/156745 | 20:34 |
| mordred | I am not going to have this particular argument in IRC since I'm going to see you this week in person | 20:34 |
| *** melwitt has joined #openstack-infra | 20:34 | |
| clarkb | or at least 1.5 because you can bake in the necessary bits | 20:34 |
| tmcpeak | fungi: ok cool, Jamie will add us now. Thank you! | 20:34 |
| *** HeOS has joined #openstack-infra | 20:34 | |
| fungi | tmcpeak: you're welcome | 20:34 |
| clarkb | mordred: if you haven't yet you should give uvtool a go | 20:35 |
| clarkb | mordred: its like docker for kvm | 20:35 |
| mordred | clarkb: neat - I'll play with it! | 20:35 |
| clarkb | and deals with all of the config drive stuff for you | 20:35 |
| krotscheck | jeblair: Wasn’t there a new “requires:” tag that you were working on? | 20:35 |
| greghaynes | clarkb: That line is amazing | 20:35 |
| clarkb | greghaynes: I knew you would like it :) | 20:35 |
| *** achanda has quit IRC | 20:36 | |
| fungi | krotscheck: are you thinking of http://lists.openstack.org/pipermail/openstack-dev/2015-February/056515.html | 20:36 |
| krotscheck | fungi: Yep. | 20:37 |
| pc_m | clarkb: ping | 20:37 |
| openstackgerrit | MORITA Kazutaka proposed openstack-infra/project-config: Add Sheepdog test job to cinder check pipeline https://review.openstack.org/153868 | 20:37 |
| clarkb | mordred: sure I will be in seattle soonish :) we can argue in person | 20:37 |
| clarkb | pc_m: hi | 20:37 |
| mordred | clarkb: :) | 20:38 |
| pc_m | clarkb: So looking for some guidance on how we should organize the directory for functional tests... | 20:38 |
| openstackgerrit | Michael Krotscheck proposed openstack-infra/puppet-storyboard: Updated storyboard configuration for pymysql. https://review.openstack.org/156755 | 20:38 |
| krotscheck | greghaynes: ^^ | 20:38 |
| pc_m | clarkb: Have the need for two functional jobs, one for each of the device drivers that VPN will have. | 20:38 |
| mordred | clarkb: do you know how to use it to boot an existing qcow2 image? | 20:38 |
| clarkb | pc_m: as I said I don't really have an opinion and the project should just decide something | 20:38 |
| clarkb | mordred: my solution for that was just copy pasta via virsh the things they do for a random qcow2 image | 20:39 |
| mordred | oh. so, it's useless then | 20:39 |
| * mordred stabs everyone | 20:39 | |
| pc_m | clarkb: understood. Just wondering what the choices are. I have some ideas, but not sure they'll work. Wondering if you (or anyone) has suggestions. | 20:39 |
| clarkb | mordred: its useful for local things | 20:39 |
| openstackgerrit | Andreas Jaeger proposed openstack-infra/infra-manual: For openstack projects, add dependency on governance change https://review.openstack.org/156745 | 20:39 |
| clarkb | mordred: its not like that iso changes much across nodes | 20:40 |
| mordred | clarkb: unless I want to use custom images | 20:40 |
| *** andreykurilin_ has quit IRC | 20:40 | |
| clarkb | s/nodes/VMs/ | 20:40 |
| clarkb | pc_m: I really don't have an opinion :) | 20:40 |
| mordred | clarkb: but all of a sudden now I'm doing crazy things with virsh | 20:40 |
| *** patrickeast has left #openstack-infra | 20:40 | |
| *** andreykurilin_ has joined #openstack-infra | 20:40 | |
| clarkb | mordred: weren't you doing those already? | 20:40 |
| mordred | GOD NO | 20:40 |
| clarkb | ah ok | 20:40 |
| mordred | I absolutely hate everything that exists about the tools for running vms on a laptop | 20:41 |
| clarkb | see I was doing carzy things with virsh completely by hand before I uvtooled | 20:41 |
| mordred | they're all terrrible | 20:41 |
| greghaynes | krotscheck: We need to either install pymysql manually there or add pymysql to requirements.txt and also wait for that patch to merge | 20:41 |
| pc_m | clarkb: fair enough. I'll try on the QA list and see if there are any suggestions. | 20:41 |
| mordred | and I havea cloud | 20:41 |
| mordred | none of them are even CLOSE to being docker | 20:41 |
| *** annegent_ has joined #openstack-infra | 20:41 | |
| krotscheck | greghaynes: I’m thinking having a default database driver packaged with storyboard is a good idea. | 20:41 |
| clarkb | mordred: greghaynes has a thing that may do what you want eventually | 20:41 |
| mordred | neat | 20:41 |
| greghaynes | clarkb: you mean my d3.sh | 20:42 |
| *** tiswanso has joined #openstack-infra | 20:42 | |
| clarkb | mordred: its basically build dib image boot that image but no config drive support yet | 20:42 |
| greghaynes | just to make js devs mad | 20:42 |
| clarkb | greghaynes: ya | 20:42 |
| krotscheck | wait what? | 20:42 |
| greghaynes | haha, its actually called dib-dev-deploy but I really wanted to call it d3.sh | 20:43 |
| fungi | your abbreviation contains an abbreviation | 20:43 |
| anteaya | greghaynes: you name stuff better than mordred | 20:45 |
| greghaynes | :) | 20:45 |
| *** dizquierdo has quit IRC | 20:46 | |
| clarkb | mordred: also maybe we should just make nova installable on its own for laptop use | 20:46 |
| clarkb | mordred: no keystone, no glance, no swift, just a nova for laptops :) | 20:46 |
| clarkb | mordred: then pip install nova-laptop just works out of the box (I can dream right) | 20:47 |
| mordred | well, nova without glance makes no sense - BUT - yes, I agree with you, I've wanted that for a while | 20:48 |
| mordred | because it turns out nova is good at launching vms without me having to edit xml | 20:48 |
| clarkb | mordred: you don't need glance because all disk images are already local right? | 20:48 |
| openstackgerrit | greghaynes proposed openstack-infra/storyboard: Add PyMySQL to requirements https://review.openstack.org/156764 | 20:48 |
| mordred | clarkb: something has to tell nova where they are | 20:48 |
| mordred | clarkb: glance is a registry | 20:48 |
| clarkb | mordred: ah I see | 20:48 |
| greghaynes | krotscheck: ^ should make your puppet edit commit message point at that | 20:48 |
| mordred | clarkb: but I still think "apt-get install nova-laptop" could still totally work | 20:49 |
| sdague | clarkb: mininova? | 20:49 |
| krotscheck | greghaynes: Oh dude, rebase that to the front. | 20:49 |
| krotscheck | Because the first patch breaks all the things. | 20:49 |
| clarkb | sdague: ya | 20:49 |
| greghaynes | krotscheck: oh? I thought we established that the first patch wont cause issues in storyboard.o.o because the puppet explicitly installs mysql-python | 20:49 |
| clarkb | sdague: that will fork off one of each necessary service and configure everything to nova networking and use sqlite | 20:49 |
| sdague | so I thought someone had done a nova api implementation like that | 20:49 |
| krotscheck | greghaynes: Yes, but we have downstream people who might be impacted. | 20:50 |
| greghaynes | psh, users | 20:50 |
| krotscheck | Who needs users. | 20:50 |
| dtroyer | https://github.com/juergh/dwarf | 20:51 |
| *** jgrimm is now known as zz_jgrimm | 20:51 | |
| openstackgerrit | Michael Krotscheck proposed openstack-infra/puppet-storyboard: Updated storyboard configuration for pymysql. https://review.openstack.org/156755 | 20:51 |
| sdague | dtroyer: ah there it is | 20:51 |
| dtroyer | I started adding vbox support to it a while back… | 20:51 |
| dtroyer | then realized that was just re-inventing nova... | 20:52 |
| mordred | :) | 20:52 |
| clarkb | ya I don't see much reason this should need to be a separate project | 20:52 |
| clarkb | its just a matter of composing an openstack nova deployment specifically for single node laptop use | 20:52 |
| *** patrickeast has joined #openstack-infra | 20:53 | |
| clarkb | this means using nova net, sqlite, and so on | 20:53 |
| openstackgerrit | Matt Riedemann proposed openstack-infra/elastic-recheck: Add query for nova resize bug 1422901 https://review.openstack.org/156766 | 20:53 |
| openstack | bug 1422901 in OpenStack Compute (nova) "test_delete_server_while_in_verify_resize_state fails with "ValueError: You must specify a valid interface name."" [High,Confirmed] https://launchpad.net/bugs/1422901 | 20:53 |
| clarkb | and people say there is no reason to support sqlite | 20:53 |
| fungi | (or nova-net!) | 20:53 |
| sdague | clarkb: so... that's interesting | 20:53 |
| *** annegent_ has quit IRC | 20:53 | |
| sdague | so if I fixture extract out the functional stuff enough, I think it's going to be pretty close to that | 20:54 |
| sdague | it would probably not take a ton to make another binary in nova that did that | 20:54 |
| *** cipcosma has quit IRC | 20:54 | |
| sdague | it's sort of nova-all, plus a glance directory | 20:55 |
| clarkb | sdague: neat | 20:55 |
| * sdague sticks it on list of things that might fall out of this | 20:55 | |
| clarkb | the number of knobs you actually need to turn are probably 1) location of image store 2) thats it go get a drink | 20:57 |
| sdague | actually, a few networking knobs as well | 20:58 |
| sdague | networking is kind of the suck | 20:58 |
| *** zz_jgrimm is now known as jgrimm | 20:58 | |
| clarkb | sdague: really? assuming its a laptop again, just pick the interface with an IP addr | 20:58 |
| *** dizquierdo has joined #openstack-infra | 20:58 | |
| clarkb | sdague: then set up your IP range in a range that doesn't conflict with that | 20:58 |
| clarkb | requires a little bit math but shouldn't be hard | 20:59 |
| clarkb | sdague: then allow for overridden things if it matters to the user | 20:59 |
| sdague | yeh, maybe, will have to ponder | 20:59 |
| sdague | I know that ends up as the sticking point inside of devstack, actually getting guests that can connect to the internet and that you can ssh to | 21:00 |
| *** unicell has quit IRC | 21:04 | |
| *** marun has quit IRC | 21:04 | |
| clarkb | sdague: so what do you need minimally? nova-api, nova-conductor, nova-compute, nova-net, nova-scheduler, somethingforglance ? | 21:04 |
| *** unicell has joined #openstack-infra | 21:04 | |
| *** unicell has quit IRC | 21:05 | |
| *** unicell1 has joined #openstack-infra | 21:05 | |
| *** unicell1 has quit IRC | 21:06 | |
| *** unicell has joined #openstack-infra | 21:06 | |
| *** hashar has joined #openstack-infra | 21:07 | |
| mordred | clarkb: you need keystone | 21:07 |
| * morganfainberg feels a disturbance in freenode. | 21:07 | |
| clarkb | mordred: hrm for some reason I thought you could just admin it up and avoid keystone with a magic string | 21:07 |
| mordred | clarkb: beacuse none of the clients do sane things without it | 21:07 |
| clarkb | oh the clients ya didnt' consider how they would be affected | 21:07 |
| mordred | keystone + nova + glance is the minimum thing you need here | 21:08 |
| clarkb | its kind of annoying that keystone is actually required | 21:08 |
| mordred | shrug | 21:08 |
| clarkb | it makes sense 99% of the time | 21:08 |
| mordred | yah - that's why it doesn't bother me :) | 21:08 |
| clarkb | mordred: well except that it means running apache on your laptop | 21:08 |
| clarkb | whihc is :( | 21:09 |
| mordred | we're talking about running libvirt and it's apache that is :( ? :) | 21:09 |
| morganfainberg | clarkb, you could run eventlet keystone | 21:09 |
| clarkb | morganfainberg: I did for a bit remember? it doesn't work :P | 21:09 |
| morganfainberg | it's not the end of the world to do that on a lap[top | 21:09 |
| * morganfainberg is missing context now... *tries to swap back into the conversation for realsies* | 21:10 | |
| clarkb | keystone eventlet is molasses slow | 21:10 |
| morganfainberg | oh. | 21:10 |
| clarkb | so the fix was well just run a shitload of them | 21:10 |
| clarkb | which isn't laptop friendly | 21:10 |
| mordred | morganfainberg: we're talking about a laptop install of openstack for ease of operating local vms | 21:10 |
| morganfainberg | right. | 21:10 |
| mordred | because all of the other tools are batshit crazy bad | 21:10 |
| morganfainberg | mordred, master or like juno? | 21:10 |
| morganfainberg | cause... next week eventlet keystone may not be that slow >.> | 21:10 |
| mordred | "it's easy, just set up your bridge and then edit this xml file with virsh..." | 21:11 |
| morganfainberg | at least that's the hope | 21:11 |
| mordred | morganfainberg: the chances we actually bother to do this are pretty low | 21:11 |
| *** hashar has quit IRC | 21:11 | |
| morganfainberg | plus using SQL Pool, and/or LDAP connection pool = 3-5x improvement | 21:11 |
| lifeless | mordred: fig + kolla | 21:11 |
| morganfainberg | makes me consider a conductor thing for keystone *wince* | 21:11 |
| lifeless | mordred: IMO | 21:11 |
| *** achanda has joined #openstack-infra | 21:11 | |
| *** achanda has quit IRC | 21:11 | |
| clarkb | morganfainberg: well the idea is this would just use sqlite, be simple | 21:12 |
| lifeless | clarkb: 'fix' :) | 21:12 |
| clarkb | lifeless: exactly :) | 21:12 |
| morganfainberg | clarkb, no don't do that pleaseeeeee don't do that | 21:12 |
| lifeless | ok, time for me to be constructive and hope on a plane | 21:12 |
| openstackgerrit | Travis McPeak proposed openstack-infra/project-config: Setting up Bandit integration with PyPI https://review.openstack.org/156772 | 21:12 |
| clarkb | morganfainberg: why not? its a laptop install | 21:12 |
| *** achanda has joined #openstack-infra | 21:12 | |
| lifeless | also sqlite is awesome | 21:12 |
| clarkb | morganfainberg: its like exactly what sqlite exists to solve | 21:12 |
| morganfainberg | sqlite doesn't really work for keystone | 21:12 |
| mordred | clarkb: mysql is not the hard part of openstack | 21:12 |
| morganfainberg | some constraint issues. | 21:12 |
| * clarkb gives up on his hopes and dreams | 21:12 | |
| *** hashar has joined #openstack-infra | 21:12 | |
| mordred | mysql on a laptop is easy and a solved problem for distro pacakges | 21:12 |
| morganfainberg | mordred, ++ | 21:12 |
| clarkb | mordred: you would never know that if you pay attention to the mail lists | 21:13 |
| clarkb | also I ran mysql on my laptop for a while (silly kde) never again | 21:13 |
| morganfainberg | clarkb, use postgres then *duckpleasedon'thurtme* | 21:13 |
| mordred | I'm trying to remember the last time there _wasn't_ a mysql | 21:13 |
| greghaynes | oh, well KDE also tried the non-server mysql for a while to | 21:13 |
| greghaynes | which was bonghits | 21:13 |
| greghaynes | "Itll be easy" they said, "itll just work" | 21:14 |
| greghaynes | lies | 21:14 |
| mordred | lifeless: fig + kolla does not even come close to solving the use case we were talking about | 21:14 |
| openstackgerrit | Merged openstack-infra/elastic-recheck: Add query for nova resize bug 1422901 https://review.openstack.org/156766 | 21:14 |
| openstack | bug 1422901 in OpenStack Compute (nova) "test_delete_server_while_in_verify_resize_state fails with "ValueError: You must specify a valid interface name."" [Critical,In progress] https://launchpad.net/bugs/1422901 - Assigned to Matt Riedemann (mriedem) | 21:14 |
| mordred | lifeless: if I want to use a container, I'll use docker, because it has an awesome UI | 21:14 |
| clarkb | there is nothing so special about these services that they couldn't all be forked and managed by some openstack-laptop parent process, all use sqlite, linux networking via nova net for comms, and zmq for message bus (avoids running another service) | 21:15 |
| clarkb | in theory all of this is already supported | 21:15 |
| clarkb | until you ever actuall look at doing anything like this then you just go back in your hole and cry | 21:15 |
| mordred | clarkb: this is why all of the vm toolchains make me so sad | 21:16 |
| mordred | clarkb: why in the hell is there not a uvt-boot $name $path_to_qcow2 | 21:16 |
| clarkb | it bugs me when people argue that "sqlite is useless kill it with fire" or "only use apache anything else isn't worth it" and so forth | 21:17 |
| *** tmcpeak has quit IRC | 21:17 | |
| openstackgerrit | Andreas Jaeger proposed openstack-infra/project-config: Bandit is a stackforge project https://review.openstack.org/156774 | 21:17 |
| *** annegent_ has joined #openstack-infra | 21:17 | |
| AJaeger | dansmith: now it's me ordering lines ^ | 21:17 |
| mordred | clarkb: except the only reason we're even having this conversation is because there is not a proper simple vm launcher | 21:18 |
| clarkb | mordred: sure I guess what I am trying to assert is that openstack isn't special here and it could/should be that too | 21:18 |
| dansmith | AJaeger: just feels right... right? | 21:18 |
| *** tmcpeak has joined #openstack-infra | 21:18 | |
| clarkb | mordred: especially since we as a project refuse to do the whole "make one single thing work well" thing | 21:18 |
| AJaeger | sometimes ;) | 21:18 |
| mordred | clarkb: it could - except our half-assed attempt to do that is what keeps us from making good architectural decisions | 21:18 |
| clarkb | mordred: if openstack said ^ instead I would definitely say don't use it for this | 21:18 |
| mordred | rather than continuing to support a toy like sqlite | 21:18 |
| mordred | if we ditched all database support except for mysql like the sane choice would be - we could actually start depending on mysql behaviors and known patterns for scale | 21:19 |
| morganfainberg | mordred, ++ i really want to kill sqlite support... heck i'd like to go out on a limb and say use mysql like XXX and it works well. | 21:19 |
| clarkb | but we do in theory support zmq which is great for this, we do in theory support sqlite which is also great for this, we do in theory support running processes without apache, and so on | 21:19 |
| mordred | as it is, we dance around with bs like "but postgres is important!!!" | 21:19 |
| mordred | it's not | 21:19 |
| mordred | a database is | 21:19 |
| morganfainberg | mordred, +∞ | 21:19 |
| clarkb | mordred: fwiw I would be fine with openstack deciding postgres was *the* database to use :P | 21:20 |
| mordred | me too | 21:20 |
| mordred | I'd be fine with either one, tbh | 21:20 |
| clarkb | but yes you need a relational DB oyu don't need 3 | 21:20 |
| morganfainberg | clarkb, a clear decision yes. | 21:20 |
| mordred | but both is stupid | 21:20 |
| mordred | and - mysql has the proven track record | 21:20 |
| *** bhuvan has joined #openstack-infra | 21:20 | |
| mordred | which is the only reason I suggest it over pg - but just pg would be 100x better than now | 21:20 |
| clarkb | mordred: its a similar story with hypervisors and message busses and so forth, some will be easier to multiplex than others | 21:21 |
| morganfainberg | mordred, honestly the only time i'd start complaining is if someone said use $non_opensource_rdbms ad the default | 21:21 |
| mordred | morganfainberg: ++ | 21:21 |
| morganfainberg | s/ad/as | 21:21 |
| mordred | clarkb: yup - see my current source of rage with rax and xen :) | 21:21 |
| mtreinish | morganfainberg: db2... :) | 21:21 |
| *** AJaeger has quit IRC | 21:21 | |
| morganfainberg | mtreinish, is is really opensource? | 21:21 |
| morganfainberg | mtreinish, if so and it works, sure. standardize | 21:22 |
| mtreinish | morganfainberg: heh, no not even close. I thought that's what you were raging about in the past | 21:22 |
| morganfainberg | mtreinish, nope. | 21:22 |
| mordred | oh - wait- I forgot - let's use mongo - that'll solve everything :) | 21:22 |
| morganfainberg | mordred, it's webscale | 21:22 |
| mordred | it's holding my coffee right now ... | 21:23 |
| mordred | so it's currently more important than any other database | 21:23 |
| *** sdake_ has joined #openstack-infra | 21:23 | |
| *** thingee has joined #openstack-infra | 21:23 | |
| morganfainberg | mordred, that is a strong argument for mongodb | 21:24 |
| morganfainberg | a very strong argument for it | 21:24 |
| morganfainberg | coffee | 21:24 |
| fungi | morganfainberg: it solves all your problems, as long as all your problems have to do with there not being enough marketing buzzwords in your documentation | 21:24 |
| thingee | overheard from morganfainberg this convo. Just wanted to pop in a say if we're considering mongodb, to rethink and consider mango db which is actually CLOUD SCALE https://github.com/dcramer/mangodb | 21:25 |
| clarkb | thingee: we already require mongodb for the most part | 21:26 |
| clarkb | several projects depend on it at this point | 21:26 |
| morganfainberg | clarkb, (it's like an onion article) | 21:26 |
| *** sdake has quit IRC | 21:27 | |
| clarkb | https://pypi.python.org/pypi/fuckitdb/0.1.3 /me hopes anteaya doesn't get mad for me linking that | 21:27 |
| clarkb | mordred: ^ you might enjoy that | 21:27 |
| mordred | clarkb: :) | 21:27 |
| fungi | "Never use this for anything important, it's an experiment." | 21:27 |
| mtreinish | clarkb: heh, I like that | 21:28 |
| morganfainberg | toss up between mangodb and that one you just linked clark | 21:28 |
| morganfainberg | both are pretty awesome. | 21:29 |
| *** annegent_ has quit IRC | 21:29 | |
| zaro | fungi: i keep getting an 'Invalid certification chain' when i try to use REST api on review-dev. do you happen to know how to make it work? | 21:29 |
| clarkb | morganfainberg: also https://github.com/philipl/pifs | 21:29 |
| fungi | zaro: how are you trying to use the rest api, specifically (from what tool/client)? | 21:29 |
| clarkb | zaro: you have to tell curl/wget/$thing to ignore untrusted certs | 21:29 |
| zaro | fungi: curl https://review-dev.openstack.org/changes/5331 | 21:29 |
| clarkb | "They said 100% compression was impossible? You're looking at it!" | 21:30 |
| clarkb | though I believe the math it relies on is unproven | 21:30 |
| fungi | zaro: curl --insecure ... | 21:30 |
| clarkb | so it may not actually work :) | 21:30 |
| fungi | zaro: according to the manpage anyway | 21:30 |
| zaro | cool. thanks. that worked. | 21:31 |
| fungi | any time | 21:31 |
| *** MaxV has joined #openstack-infra | 21:32 | |
| clarkb | jhesketh: ok I think I am up to date reviewing the various swift related changes | 21:32 |
| clarkb | I have also reviewed the specs that were pointed out | 21:32 |
| clarkb | asselin: ^ I left comments on yours | 21:32 |
| clarkb | jhesketh: also sigmavirus24 says we should use threads/processes and do post one file at a time using requests.post | 21:33 |
| sigmavirus24 | Well you're currently posting a single file at a time serially | 21:33 |
| clarkb | sigmavirus24: right | 21:33 |
| sigmavirus24 | You can do the same in threads safely | 21:33 |
| clarkb | sigmavirus24: serially? | 21:33 |
| *** changbl has quit IRC | 21:33 | |
| sigmavirus24 | In each thread, yes | 21:34 |
| clarkb | kk | 21:34 |
| sigmavirus24 | Unless there's a different way to upload multiple files simultaneously to swift | 21:34 |
| clarkb | I need food then will likely hack on that next | 21:34 |
| notmyname | ? | 21:34 |
| clarkb | sigmavirus24: I think thats the way to do it, iirc the swiftclient implementation doesn't work though | 21:34 |
| clarkb | notmyname: uploading >1 file to swift concurrently | 21:35 |
| *** mwagner_lap has quit IRC | 21:35 | |
| notmyname | clarkb: yes. do that to saturate your network | 21:35 |
| notmyname | sigmavirus24: clarkb: any chance you have the files locally in a .tar[.gz]? | 21:36 |
| * sigmavirus24 is just advising on the usage of requests | 21:36 | |
| sigmavirus24 | I think the log files are usually compressed before hand though | 21:37 |
| *** koolhead17 has quit IRC | 21:37 | |
| clarkb | they are compressed before hand but not tarred, and if we added tarballs we would likely still need them exploded out too | 21:37 |
| notmyname | swift has a feature to allow you to upload a .tar[.gz] file and have the components stored as separate objects in the cluster | 21:37 |
| clarkb | notmyname: oh! that might actually be handy, is the tar still available too? | 21:38 |
| jogo | Releases are named after Van Halen songs. | 21:38 |
| notmyname | but that doesn't help the speed. they'd still be uploaded serially as the tar is streamed | 21:38 |
| *** changbl has joined #openstack-infra | 21:38 | |
| jogo | ansible has a good naming scheme | 21:38 |
| notmyname | clarkb: no. | 21:38 |
| clarkb | notmyname: ok so likely wejust want to do each file upload in a different thread and let the IO saturate a bit more | 21:38 |
| *** e0ne has quit IRC | 21:39 | |
| notmyname | clarkb: yes. that's the normal best practice | 21:40 |
| sigmavirus24 | clarkb: notmyname just to be clear, the way you're currently doing this, you're reading the entire file into memory, storing it in a string and sending that all at once to swift | 21:41 |
| notmyname | sigmavirus24: no. that would be terrible :-) | 21:41 |
| sigmavirus24 | If you were streaming the file upload it would take forever | 21:41 |
| sigmavirus24 | notmyname: no that's what is currently happening :D | 21:41 |
| clarkb | having the file in memory should be fine right? this is about network IO not memory/disk | 21:42 |
| notmyname | oh. then I'm not sure. like you're helping with requests, i'm the guy who does the swift backend part. i'm not sure of the infra tooling :-) | 21:42 |
| sigmavirus24 | clarkb: okay just making sure we were on the same page :) | 21:42 |
| sigmavirus24 | notmyname: I'm one of the requests core developers in case you didn't already know | 21:42 |
| clarkb | sigmavirus24: I mean the kernel is gonna cache it all in memory anyways so this shouldn't be much different | 21:42 |
| sigmavirus24 | clarkb: :thumbsup: | 21:43 |
| notmyname | sigmavirus24: ya, I figured that out :-) | 21:43 |
| sigmavirus24 | I keep noodling on an idea to improve streamed file uploads but I keep thinking it's a bad bad idea that should never be done :-P | 21:43 |
| sigmavirus24 | (Actually 2 ideas) | 21:43 |
| *** rm_work is now known as rm_work|away | 21:44 | |
| fungi | the only thing better than a bad idea is 2 bad ideas | 21:44 |
| clarkb | fungi: 3 bad ideas make a right idea | 21:45 |
| sdague | notmyname: so regarding the earlier patch, swiftclient doesn't seem to have the tox target, can you take a look at my change to add it - https://review.openstack.org/#/c/156719/ so we can keep the ball rolling? | 21:48 |
| *** ociuhandu has joined #openstack-infra | 21:48 | |
| notmyname | sdague: what's up with http://logs.openstack.org/19/156719/1/check/check-swiftclient-dsvm-functional/ac76eb3/console.html#_2015-02-17_19_42_33_029 | 21:48 |
| sdague | notmyname: all check-swiftclient-dsvm-functional fail like that | 21:49 |
| notmyname | sdague: ok let me rephrase. why? ;-) | 21:49 |
| *** otter768 has joined #openstack-infra | 21:49 | |
| sdague | does that jobs specify swiftclient installs from git? | 21:50 |
| notmyname | I don't know | 21:50 |
| sdague | that's my guess | 21:50 |
| sdague | because by default devstack now installs libraries from pypi, not from git trees | 21:51 |
| sdague | that changed back in ... Nov? | 21:51 |
| *** changbl has quit IRC | 21:51 | |
| *** changbl has joined #openstack-infra | 21:52 | |
| openstackgerrit | Sean Dague proposed openstack-infra/project-config: make swiftclient functional job use swiftclient from git https://review.openstack.org/156787 | 21:54 |
| sdague | notmyname: I think that ^^^ will fix it | 21:54 |
| notmyname | ok | 21:54 |
| *** otter768 has quit IRC | 21:54 | |
| *** gyee has quit IRC | 21:55 | |
| openstackgerrit | Clark Boylan proposed openstack-infra/project-config: Upload files concurrently to swift. https://review.openstack.org/156788 | 21:56 |
| clarkb | so that change is completely untested :) I am going to WIP it until I get a chance to test | 21:56 |
| clarkb | jhesketh: ^ but I think that should help | 21:57 |
| jeblair | clarkb: is the short version of the outcome that individual files are currently as fast as they can be and we should simply upload all files in parallel? | 21:58 |
| openstackgerrit | Merged openstack-infra/storyboard: Fix comments put method https://review.openstack.org/155790 | 21:58 |
| clarkb | jeblair: yes notmyname suggests not uploading serially :) | 21:58 |
| clarkb | jeblair: the old code uploaded 10 files at a time serially | 21:58 |
| notmyname | clarkb: jeblair: normally uploading a single object to swift will not saturate your uplink. so uploading a bunch at once makes it go faster | 21:59 |
| notmyname | clarkb: jeblair: if that doesn't work, we could consider adding racing stripes and a spoiler | 21:59 |
| jeblair | notmyname: makes sense; just wanted to verify that we'd determined that's the bottleneck | 22:00 |
| jeblair | clarkb: 10 files? | 22:00 |
| *** bknudson has quit IRC | 22:00 | |
| sigmavirus24 | jeblair: I can confirm, it was happening serially and it was killing performance | 22:01 |
| clarkb | jeblairs ya pretty sure it iterates 10 at a time | 22:01 |
| * sigmavirus24 always shoots for the puns | 22:01 | |
| jeblair | clarkb: i don't quite understand serially 10 at a time | 22:01 |
| BobBall_AWOL | How much memory should zuul-server take? is 3gb normal or unreasonable? ;) | 22:02 |
| *** marun has joined #openstack-infra | 22:02 | |
| clarkb | jeblair that is what it was doing | 22:02 |
| clarkb | it gave requests a dict of 10 files to post. requests does so serially | 22:03 |
| *** Hefeweizen has quit IRC | 22:03 | |
| sigmavirus24 | wait a second. I misunderstood the code | 22:05 |
| jeblair | clarkb: where's 10 come from? | 22:05 |
| jeblair | clarkb: like, what varable has a value of 10? | 22:05 |
| sigmavirus24 | clarkb: https://git.openstack.org/cgit/openstack-infra/project-config/tree/jenkins/scripts/zuul_swift_upload.py#n174 is sending 10 files at once? If that's the case it's sending them all at once | 22:07 |
| sigmavirus24 | i.e. files has 10 files in it? | 22:07 |
| clarkb | jeblair: https://git.openstack.org/cgit/openstack-infra/project-config/tree/jenkins/scripts/zuul_swift_upload.py#n151 from there | 22:07 |
| clarkb | jeblair: I think its a thing that zuul sets | 22:07 |
| clarkb | sigmavirus24: yes | 22:08 |
| sdague | fungi: https://review.openstack.org/#/c/156676/ the depends on won't help right, because they don't share a queue? | 22:08 |
| clarkb | sigmavirus24: but its doing so as a serial stream like a tarball right? | 22:08 |
| sigmavirus24 | I misunderstood that then. It's doing them all as one giant multipart/form-data post | 22:08 |
| clarkb | ya ok | 22:08 |
| fungi | sdague: it will still prevent one from merging before the other | 22:08 |
| sdague | fungi: it will? | 22:08 |
| sigmavirus24 | So it reads all of them into a string and send that to the server | 22:08 |
| clarkb | so I think this should still help things but maybe my change isn't quite right | 22:08 |
| sdague | I didn't realize that | 22:08 |
| fungi | sdague: even if they don't share a queue, yes | 22:08 |
| sdague | ok, great, will do | 22:08 |
| openstackgerrit | greghaynes proposed openstack-infra/storyboard: Switch from mysql-python to PyMySQL https://review.openstack.org/152339 | 22:08 |
| openstackgerrit | Sean Dague proposed openstack-infra/project-config: have swift functional tests use tox https://review.openstack.org/156676 | 22:09 |
| openstackgerrit | Sean Dague proposed openstack-infra/project-config: make swiftclient functional job use swiftclient from git https://review.openstack.org/156787 | 22:09 |
| openstackgerrit | greghaynes proposed openstack-infra/storyboard: Reorder py27 test to happen after py3 tests https://review.openstack.org/155161 | 22:09 |
| sdague | and I get to use the magic rebase button \o/ | 22:09 |
| openstackgerrit | greghaynes proposed openstack-infra/storyboard: Use python3 compatible httpclient https://review.openstack.org/155139 | 22:09 |
| openstackgerrit | greghaynes proposed openstack-infra/storyboard: Use python3 compatible unicode types https://review.openstack.org/155140 | 22:09 |
| openstackgerrit | greghaynes proposed openstack-infra/storyboard: Use python3 compatible urllib https://review.openstack.org/155150 | 22:10 |
| openstackgerrit | greghaynes proposed openstack-infra/storyboard: Use Python 3 compatible configparser https://review.openstack.org/155922 | 22:10 |
| fungi | sdague: basically referring to the gate pipeline section of jeblair's announcement http://lists.openstack.org/pipermail/openstack-dev/2015-February/056515.html which addresses that case specifically | 22:10 |
| openstackgerrit | greghaynes proposed openstack-infra/storyboard: Use TempDir fixture for working dir https://review.openstack.org/155956 | 22:10 |
| fungi | (in the second paragraph) | 22:10 |
| *** Hefeweizen has joined #openstack-infra | 22:10 | |
| sdague | fungi: ok, I guess I read too fast | 22:10 |
| openstackgerrit | greghaynes proposed openstack-infra/storyboard: Fix bytes vs string python 3 issues https://review.openstack.org/155969 | 22:10 |
| *** achanda has quit IRC | 22:10 | |
| *** ghostpl_ has joined #openstack-infra | 22:10 | |
| *** salv-orlando has quit IRC | 22:10 | |
| fungi | sdague: makes for an excellent mechanically-enforced safety net | 22:10 |
| sdague | yep | 22:11 |
| * sigmavirus24 just noticed that storyboard is using a GPL'd library | 22:11 | |
| krotscheck | Which one? | 22:11 |
| *** vilobhmm has quit IRC | 22:11 | |
| *** hashar has quit IRC | 22:11 | |
| sdague | ok, good, so that's all queued up and just waiting for the swift team to approve the tox change - https://review.openstack.org/#/c/156719/ | 22:12 |
| *** crc32 has joined #openstack-infra | 22:12 | |
| *** changbl has quit IRC | 22:12 | |
| *** vilobhmm has joined #openstack-infra | 22:13 | |
| *** salv-orlando has joined #openstack-infra | 22:13 | |
| *** erikmwilson has quit IRC | 22:13 | |
| *** erikmwil- is now known as erikmwilson | 22:13 | |
| *** mugsie has quit IRC | 22:13 | |
| sigmavirus24 | krotscheck: rfc3987 is GPL'd | 22:13 |
| *** mugsie has joined #openstack-infra | 22:14 | |
| sigmavirus24 | If you're looking for RFC 3986 compatible URI parsing/validation then rfc3986 has it and is already in g-r | 22:14 |
| sigmavirus24 | rfc3986 is also Apache 2.0 | 22:14 |
| krotscheck | sigmavirus24: I’ll look into it. | 22:14 |
| sigmavirus24 | krotscheck: different API though probably and Nova's using it for URI validation (or maybe keystone is, I forget which service) | 22:15 |
| krotscheck | sigmavirus24: python-crontab is also gpl'd | 22:15 |
| *** ghostpl_ has quit IRC | 22:15 | |
| openstackgerrit | Merged openstack-infra/jenkins-job-builder: Fix wrong casing in JDKAxis <name> tag https://review.openstack.org/156519 | 22:15 |
| krotscheck | sigmavirus24: As is pika | 22:15 |
| jhesketh | clarkb: (catching up over here). Cool, that looks good, although we're still reading all the files into memory | 22:15 |
| *** jedimike has quit IRC | 22:16 | |
| stevemar | rodrigods, around? | 22:16 |
| clarkb | jhesketh: I don't think reding the files to memory is a big deal they end up in memory regardless | 22:16 |
| jhesketh | true | 22:16 |
| marun | we have a funny problem with tox right now | 22:16 |
| stevemar | oops wrong channel | 22:16 |
| *** unicell has quit IRC | 22:16 | |
| fungi | marun: knee-slapping funny or side-splitting funny? i have to prepare myself accordingly | 22:16 |
| clarkb | krotscheck: apscheduler for cron like things is what zuul uses | 22:16 |
| clarkb | krotscheck: and pika's package page claims to be mozilla public licensed | 22:17 |
| marun | fungi: sad funny? | 22:17 |
| marun | https://review.openstack.org/#/c/156767/3/tox.ini | 22:17 |
| *** unicell has joined #openstack-infra | 22:17 | |
| fungi | aww, that's less fun | 22:17 |
| openstackgerrit | Max Rasskazov proposed openstack-infra/jenkins-job-builder: Add support of multiple ssh-agent credentials https://review.openstack.org/156681 | 22:17 |
| krotscheck | clarkb: It’s dual licenced. | 22:17 |
| marun | fungi: less fun, indeed. | 22:17 |
| clarkb | krotscheck: then you can ignore that it is gpl if you use it under the terms of the other license | 22:17 |
| marun | fungi: the syntax on the left requires tox > 1.8, but 1404 ships with 1.6 | 22:17 |
| sigmavirus24 | krotscheck: rfc3987 is GPL GPL'd, not LGPL'd or AGPL'd though which I thought was where we drew the line (I'm not bothered by it personally) | 22:18 |
| *** crc32 has quit IRC | 22:18 | |
| marun | fungi: we're wondering whether to set minversion and force developers to figure out how to install a new-enough version, or use the syntax on the right for distro package compat | 22:18 |
| * marun shakes fist at distro-packaged tox | 22:18 | |
| *** Mmike has quit IRC | 22:18 | |
| clarkb | marun: fungi its for development, just set minver imo | 22:18 |
| fungi | marun: i guess it depends on what releases of tox you're supporting. we install the latest tox, pip, virtualenv, setuptools, et cetera on our workers | 22:19 |
| marun | fungi: latest from pypi? | 22:19 |
| fungi | marun: yep | 22:19 |
| marun | I guess we'll go with minver then | 22:19 |
| clarkb | you also needed newer tox than trusty has to support py34 | 22:19 |
| krotscheck | sigmavirus24: I am. I think the GPL needs to die in a fire. | 22:19 |
| marun | clarkb: ah, that settles it then | 22:19 |
| fungi | and i run the latest pypi version on my dev systems/workstations too | 22:19 |
| clarkb | marun: you need 1.7.0 for that iirc | 22:19 |
| sigmavirus24 | krotscheck: fair enough | 22:19 |
| sigmavirus24 | I'm not so opposed to the GPL but I know lots are which is why I pointed it out | 22:20 |
| marun | clarkb, fungi: danke! | 22:20 |
| fungi | krotscheck: that's too bad. i don't use gpl for my personal projects, but it's definitely got some good points and has done a lot for software freedom in general | 22:20 |
| * greghaynes throws a wet rms at krotscheck | 22:20 | |
| krotscheck | fungi: You’re entitled to your opinion, as am I. | 22:20 |
| *** melwitt has quit IRC | 22:21 | |
| clarkb | the wet rms falls over then gets on a bus in the wrong direction | 22:21 |
| sigmavirus24 | greghaynes++ | 22:21 |
| *** Mmike has joined #openstack-infra | 22:21 | |
| *** melwitt has joined #openstack-infra | 22:21 | |
| greghaynes | clarkb: oh no, not again | 22:22 |
| sigmavirus24 | At least wet rms didn't reject clang improvements to gcc | 22:22 |
| greghaynes | (nibalizer may have actually sent an rms on a bus in the wrong direction when he came to visit us at uni) | 22:22 |
| sigmavirus24 | krotscheck: if you remind me some other time I'll try to shim rfc3986 into storyboard for you or just build in the API y'all need for it to work well | 22:24 |
| *** thingee has quit IRC | 22:25 | |
| krotscheck | sigmavirus24: I’m working on it. | 22:26 |
| sigmavirus24 | :thumbsup: | 22:26 |
| *** melwitt has quit IRC | 22:27 | |
| clarkb | jhesketh: I think my biggest current concern with that is going to be api rate limiting | 22:27 |
| *** dkranz has quit IRC | 22:27 | |
| clarkb | jhesketh: which was potentially a problem before but my change means we will hit it 10x quicker | 22:27 |
| jhesketh | clarkb: yeah I hadn't thought of that... | 22:28 |
| *** achanda has joined #openstack-infra | 22:29 | |
| *** rverchikov has left #openstack-infra | 22:30 | |
| jhesketh | notmyname: maybe you can save me some time here.. we were using swift's formpost middleware because I thought I could push x (10) files at at time rather than with the tempurl where we can only do one... If we are threading each form post individually anyway, should we instead go back to tempurl | 22:30 |
| jhesketh | (the downside with tempurl is that we need to know the full path to push into beforehand so it's likely that's not practical) | 22:31 |
| *** MaxV has quit IRC | 22:31 | |
| *** crc32 has joined #openstack-infra | 22:31 | |
| *** amitgandhinz has quit IRC | 22:31 | |
| notmyname | jhesketh: slight diversion on tempurl, but just last week we added container-level tempurl keys. so that's cool and something to look forward to | 22:31 |
| *** erikwilson has joined #openstack-infra | 22:32 | |
| notmyname | jhesketh: but wouldn't formpost still do the requests serially? it's just on POST with a huge MIME payload, right? | 22:32 |
| notmyname | jhesketh: yes, it's 10 files "at one time". but more specifically, it's 10 files in one request. not concurrently | 22:32 |
| clarkb | notmyname: right its similar to your tar example | 22:33 |
| notmyname | yup | 22:33 |
| clarkb | basically everything is lined up as one long string then copied sequentially | 22:33 |
| *** bcourt has quit IRC | 22:33 | |
| notmyname | yes | 22:33 |
| jhesketh | yeah, thought as such, just checking your opinion on the difference | 22:34 |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/project-config: Add a project to track DefCore reference data https://review.openstack.org/156807 | 22:36 |
| fungi | hogepodge: ^ | 22:36 |
| jhesketh | clarkb: so in your change we're still looping over the files in chunks set by the max_file_count out of the formpost middleware. ie, we're doing 10 threads at a time, waiting for those to finish, then starting another 10 threads | 22:37 |
| clarkb | jhesketh: yes | 22:37 |
| clarkb | jhesketh: I figured it needed some sort of rate limiting | 22:37 |
| fungi | hogepodge: the split-out repo clone url is in the gerrit/projects.yaml addition. clone a copy of that and make sure it's how you want | 22:37 |
| *** tiswanso has quit IRC | 22:37 | |
| jhesketh | clarkb: would it be better to create a thread pool? | 22:37 |
| jhesketh | then we're not waiting on the really large files to finish before starting the next batch | 22:38 |
| cebruns | What's the best way to request a new feature for Gerrit code review flow? Basically would like style checks failures to be inlined as code review comments rather than having to dig through logs. | 22:38 |
| *** melwitt has joined #openstack-infra | 22:39 | |
| *** patrickeast has quit IRC | 22:39 | |
| jeblair | krotscheck: i believe that use of plain gpl (not lgpl) libraries by liberally licensed python libraries is unsettled territory, but potentially problematic, and should probably be avoided. | 22:40 |
| jeblair | krotscheck: i believe that use of plain gpl (not lgpl) libraries by liberally licensed python programs is unsettled territory, but potentially problematic, and should probably be avoided. | 22:40 |
| jeblair | correction ^ | 22:40 |
| clarkb | jhesketh: ya thats a lot less simple though :) | 22:41 |
| krotscheck | jeblair: Yep. Working on that. | 22:41 |
| jeblair | and in the future, it would be great if we could talk about this without using charged language. | 22:41 |
| nibalizer | yes my first guidline is not to play host for rms | 22:42 |
| clarkb | jhesketh: but if you think that would be a good improvement I can work on it | 22:42 |
| nibalizer | my second guideline is if you ignored #1, don't lose him | 22:42 |
| *** patrickeast has joined #openstack-infra | 22:42 | |
| *** rlandy has quit IRC | 22:42 | |
| jhesketh | clarkb: any thoughts how to test it? | 22:44 |
| *** sdake_ws has quit IRC | 22:44 | |
| *** amitgandhinz has joined #openstack-infra | 22:44 | |
| clarkb | jhesketh: I was planning on maybe grabbing the files from swift for a job that ran the upload them again to a different location and see if it goes any quicker | 22:45 |
| clarkb | jhesketh: but I haven't start that process | 22:45 |
| *** andreykurilin_ has quit IRC | 22:45 | |
| *** pc_m has quit IRC | 22:46 | |
| *** bknudson has joined #openstack-infra | 22:48 | |
| krotscheck | jeblair: What happens if a library that we’re using makes use of a GPL library? | 22:48 |
| * krotscheck is looking at jsonschema, which uses rfc3987 | 22:48 | |
| krotscheck | (The first is MIT, the second GPL) | 22:48 |
| *** tmcpeak has quit IRC | 22:48 | |
| openstackgerrit | Matt Riedemann proposed openstack-infra/elastic-recheck: Update query for bug 1422901 https://review.openstack.org/156819 | 22:49 |
| *** tmcpeak has joined #openstack-infra | 22:49 | |
| openstack | bug 1422901 in OpenStack Compute (nova) "test_delete_server_while_in_verify_resize_state fails with "ValueError: You must specify a valid interface name."" [Critical,In progress] https://launchpad.net/bugs/1422901 - Assigned to Matt Riedemann (mriedem) | 22:49 |
| clarkb | jhesketh: I will probably take a stab at the thread pool first because you are right that it will be better | 22:49 |
| *** gyee has joined #openstack-infra | 22:49 | |
| jeblair | krotscheck: yeesh! that's confusing. i think the safe thing would be to avoid it | 22:49 |
| krotscheck | Yeah, I’m going to file a ticket on that github project. | 22:49 |
| jeblair | krotscheck: it's interesting to note jsonschema is in global-requirements | 22:50 |
| mtreinish | jeblair: yeah jsonschema is being pretty widely used | 22:51 |
| mtreinish | jeblair: both nova and tempest have been using it for some time now | 22:51 |
| mtreinish | glance also | 22:51 |
| *** melwitt has quit IRC | 22:51 | |
| *** ivar-laz_ has joined #openstack-infra | 22:52 | |
| jhesketh | clarkb: that's a good idea (re re-uploading). If we can get a job held we should still have the same creds from zuul | 22:52 |
| jhesketh | clarkb: do you want me to help with the thread pool? | 22:52 |
| krotscheck | yeaaahhh.... | 22:52 |
| *** camunoz_gone is now known as camunoz | 22:53 | |
| sigmavirus24 | krotscheck: jeblair so jsonschema is but rfc3987 is optional and rfc3986 has shim code in nova (or keystone) for the validation that's performed | 22:53 |
| clarkb | jhesketh: nah I think I have it, just going to put a bunch of (url, kwarg) tuples into a queue, then start X threads that exit when queue is empty | 22:53 |
| sigmavirus24 | clarkb: does multiprocessing.pool.ThreadPool not exist on Py3? | 22:53 |
| sigmavirus24 | *Py2 | 22:53 |
| *** melwitt has joined #openstack-infra | 22:53 | |
| *** patrickeast has quit IRC | 22:53 | |
| krotscheck | sigmavirus24 jeblair: Yeah, there’s a try/catch block around the import | 22:54 |
| jhesketh | clarkb: sounds good | 22:54 |
| *** doug-fish has left #openstack-infra | 22:54 | |
| *** mfink_ has quit IRC | 22:54 | |
| sigmavirus24 | krotscheck: yeah, rfc3986 is only used by one openstack service now for URI validation but it can hook into jsonschema as necessary | 22:54 |
| * sigmavirus24 should probably add something to rfc3986 that just auto-registers itself as a validator | 22:55 | |
| *** ivar-lazzaro has quit IRC | 22:55 | |
| clarkb | sigmavirus24: neither python2 nor python 3 documentation document that object | 22:55 |
| sigmavirus24 | clarkb: intersting | 22:55 |
| sigmavirus24 | *interesting, I'm almost certain I've used that | 22:55 |
| clarkb | also I was pretty sure multiprocessing did processes | 22:56 |
| sigmavirus24 | clarkb: so ThreadPool is for using the threading library iirc | 22:57 |
| *** patrickeast has joined #openstack-infra | 22:57 | |
| sigmavirus24 | multiprocessing.Pool is the process-based pool iirc | 22:57 |
| * sigmavirus24 could be mis-remembering though | 22:59 | |
| *** ghostpl_ has joined #openstack-infra | 23:00 | |
| *** dustins_ has quit IRC | 23:02 | |
| *** melwitt has quit IRC | 23:02 | |
| cebruns | jeblair: What's the best way to request a new feature for Gerrit code review flow? Cinder folks looking for an enhancement. | 23:04 |
| jeblair | zaro: ^ | 23:04 |
| jeblair | cebruns: they have an issue tracker and mailing list... | 23:04 |
| jeblair | cebruns: probably file an issue if you just want to record the suggestion... | 23:05 |
| jeblair | cebruns: if you have resources to develop it, i'd start talking on their mailing list | 23:05 |
| *** ociuhandu has quit IRC | 23:06 | |
| cebruns | jeblair: Great - thanks! | 23:06 |
| *** dizquierdo has quit IRC | 23:06 | |
| *** thedodd has joined #openstack-infra | 23:07 | |
| fungi | cebruns: http://code.google.com/p/gerrit/issues/list | 23:07 |
| *** zxiiro has quit IRC | 23:07 | |
| *** dkliban is now known as dkliban_afk | 23:08 | |
| *** esker has quit IRC | 23:08 | |
| *** chlong has joined #openstack-infra | 23:08 | |
| *** vigneshvar has quit IRC | 23:09 | |
| *** mriedem has quit IRC | 23:09 | |
| cebruns | fungi jeblair: Thanks - I was thinking more an infra script that would run a style check test/review for any commits (see https://etherpad.openstack.org/p/styleChecksInReviews) Would that be a BP request for Infra? | 23:09 |
| *** zxiiro has joined #openstack-infra | 23:10 | |
| *** ddieterly has quit IRC | 23:10 | |
| *** hichihara has joined #openstack-infra | 23:10 | |
| jeblair | cebruns: that's probably something that should be done in zuul | 23:10 |
| *** ddieterly has joined #openstack-infra | 23:11 | |
| *** ddieterly has quit IRC | 23:13 | |
| *** ddieterl_ has joined #openstack-infra | 23:13 | |
| cebruns | jeblair: Great - would I submit a BP to get the feature request noted/tracked? | 23:15 |
| openstackgerrit | Darragh Bailey proposed openstack-infra/jenkins-job-builder: Support excludes when recursively processing paths https://review.openstack.org/68895 | 23:15 |
| openstackgerrit | Michael Krotscheck proposed openstack-infra/storyboard: Removed rfc3987 library https://review.openstack.org/156833 | 23:16 |
| jeblair | cebruns: you can file a story here: https://storyboard.openstack.org/#!/project/679 | 23:16 |
| jeblair | cebruns: i think actual implementation of that would require an infra-spec | 23:16 |
| jeblair | cebruns: because it will require some coordination between zuul and the jobs that it runs | 23:16 |
| clarkb | would need to serialize the lint output to someting that zuul could process right? | 23:17 |
| jeblair | clarkb: yeah, that's what i'm thinking | 23:17 |
| jeblair | clarkb: and then need to be handed back to zuul somehow | 23:17 |
| jeblair | then zuul can post to gerrit | 23:17 |
| jeblair | cebruns: is anyone interested in hacking on this? would be mostly python i think. | 23:17 |
| cebruns | jeblair: nobody tagged right now - but it would be really useful. Is it one of those "whoever brought it up...." type of assignments? :) | 23:19 |
| *** timcline has quit IRC | 23:21 | |
| jeblair | cebruns: i think anyone could pick it up, but our backlog is pretty deep, so if someone is enthusastic enough to start hacking on it, i'd be happy to help them get started | 23:21 |
| *** yamamoto has joined #openstack-infra | 23:21 | |
| anteaya | clarkb: /me is disillusioned and could care less | 23:23 |
| cebruns | jeblair: I'm willing to give it a go - just need some direction to get going. Thx! | 23:23 |
| jeblair | cebruns: okay, so why don't you send a mail to the openstack-infra@ mailing list with the suggestion, and we can respond with some suggestions about how to do it | 23:24 |
| *** tmcpeak has quit IRC | 23:24 | |
| jeblair | cebruns: that should get us in place for a short spec | 23:24 |
| *** melwitt has joined #openstack-infra | 23:25 | |
| *** signed8bit has quit IRC | 23:25 | |
| cebruns | jeblair: Will do. | 23:26 |
| openstackgerrit | Darragh Bailey proposed stackforge/python-jenkins: Provider helper classes for plugin data management https://review.openstack.org/154316 | 23:26 |
| *** gordc has quit IRC | 23:26 | |
| openstackgerrit | Merged openstack-infra/elastic-recheck: Update query for bug 1422901 https://review.openstack.org/156819 | 23:27 |
| openstack | bug 1422901 in OpenStack Compute (nova) "test_delete_server_while_in_verify_resize_state fails with "ValueError: You must specify a valid interface name."" [Critical,In progress] https://launchpad.net/bugs/1422901 - Assigned to Matt Riedemann (mriedem) | 23:27 |
| *** pradk has quit IRC | 23:29 | |
| *** yamamoto has quit IRC | 23:29 | |
| *** dmsimard is now known as dmsimard_away | 23:30 | |
| *** kgiusti has quit IRC | 23:30 | |
| jogo | clarkb: any luck reproducing the aiopcpu issue around the subnode? | 23:31 |
| zaro | jeblair: trying to use gertty with review-dev but running into issues. | 23:31 |
| openstackgerrit | Clark Boylan proposed openstack-infra/project-config: Upload files concurrently to swift. https://review.openstack.org/156788 | 23:31 |
| clarkb | jogo: no | 23:31 |
| clarkb | jhesketh: ^ still completely untested but that should be at least more structurally sound | 23:31 |
| zaro | jeblair: looks like gertty always attempts to clone with '/p/' in url which won't work if gerrit isn't setup with the reverse proxy thing correct? | 23:32 |
| jogo | clarkb: hmm | 23:32 |
| *** dimtruck is now known as zz_dimtruck | 23:32 | |
| jogo | well tempest-full multi node is working now | 23:32 |
| clarkb | jogo: but not with neutron right? seems like I saw neutron failing but nova net worked | 23:33 |
| zaro | jeblair: or maybe it's the rewrite rules. | 23:33 |
| *** tkelsey has quit IRC | 23:33 | |
| clarkb | zaro: no gerrit supports the /p/ url out of the box | 23:33 |
| jhesketh | clarkb: cool, will look :-) | 23:34 |
| zaro | clarkb: does it need to be enabled or something? | 23:34 |
| clarkb | zaro: no I don't think so, it was just stanard | 23:34 |
| zaro | clarkb: it's not available for review-dev | 23:34 |
| *** ddieterl_ has quit IRC | 23:35 | |
| zaro | clarkb: i see it in gerrit.vhost.erb file in puppet-gerrit repo. | 23:35 |
| *** ddieterly has joined #openstack-infra | 23:35 | |
| clarkb | zaro: that may be because our apache vhost config doesn't work with apache 2.4 | 23:35 |
| clarkb | seems like asselin had to fix it for the zuul mergers to run on trusty | 23:35 |
| asselin | hi | 23:36 |
| clarkb | hey there I thinkzaro needs your apache2.4 magic for gerrit repo mirrors | 23:37 |
| zaro | clarkb: i don't see any reference to /p/ in gerrit docs. what would it be under? | 23:37 |
| asselin | I had to add some additions: https://review.openstack.org/#/c/155154/ | 23:37 |
| fungi | yeah, the cgi perms | 23:37 |
| fungi | most likely | 23:37 |
| clarkb | zaro try removing the stuff in apache in review-dev it should work via gerrit then | 23:37 |
| clarkb | but we want to fix the apache config | 23:37 |
| zaro | will give that a try. thanks | 23:39 |
| clarkb | cebruns fwiw it would be super nice if devs always ran the linter(s) locally and those jobs never failed :) | 23:39 |
| jogo | clarkb: hmm let me double check I think neutron actually works now | 23:40 |
| cebruns | clarkb: Yep - I agree with ya. | 23:41 |
| *** MaxV has joined #openstack-infra | 23:41 | |
| *** MaxV has quit IRC | 23:46 | |
| marun | jeblair: did you get a chance to look at my naive bugfix? | 23:48 |
| *** ivar-laz_ has quit IRC | 23:48 | |
| jhesketh | clarkb: commented on 156788. Just one small thing and it lgtm, the rest are nits | 23:48 |
| marun | jeblair: https://review.openstack.org/#/c/156716/ | 23:49 |
| *** ivar-lazzaro has joined #openstack-infra | 23:49 | |
| zaro | clarkb: there's lines for 'AliasMatch', 'ScriptAlias', and 'RewriteCond' in gerrit.vhost.erb should all of those be removed? | 23:49 |
| clarkb | jhesketh: thanks and good catch, ya I kept the whole fileincr thing just so that multiple files per post were still a possibility | 23:49 |
| marun | jeblair: the test failure I see in tests.test_scheduler.TestScheduler.test_client_get_running_jobs suggests that some metajobs may actually depend on voting defaulting to true? | 23:50 |
| *** otter768 has joined #openstack-infra | 23:50 | |
| marun | jeblair: I was assuming that metajobs should not be setting any values that aren't explicitly defined, is that incorrect? | 23:50 |
| *** signed8bit has joined #openstack-infra | 23:52 | |
| *** tmcpeak has joined #openstack-infra | 23:52 | |
| *** yamamoto has joined #openstack-infra | 23:52 | |
| openstackgerrit | Joe Gordon proposed openstack-infra/devstack-gate: Set up ssh_known_host based on hostname https://review.openstack.org/136596 | 23:52 |
| openstackgerrit | Joe Gordon proposed openstack-infra/devstack-gate: Set custom cpu_model for live_migrate https://review.openstack.org/141530 | 23:52 |
| openstackgerrit | Joe Gordon proposed openstack-infra/devstack-gate: Setup /etc/hosts for aiopcpu https://review.openstack.org/136158 | 23:52 |
| openstackgerrit | Joe Gordon proposed openstack-infra/devstack-gate: Allow root user to ssh in as stack https://review.openstack.org/137176 | 23:52 |
| openstackgerrit | Joe Gordon proposed openstack-infra/devstack-gate: Enable live block migration https://review.openstack.org/135768 | 23:52 |
| *** ZZelle_ has quit IRC | 23:53 | |
| jeblair | marun: not yet, sorry | 23:53 |
| *** abhirc has quit IRC | 23:53 | |
| marun | jeblair: no worries | 23:54 |
| *** amitgandhinz has quit IRC | 23:54 | |
| jogo | clarkb: moving on to getting block migration working | 23:54 |
| *** otter768 has quit IRC | 23:55 | |
| *** tmcpeak has quit IRC | 23:55 | |
| zaro | clarkb: i removed all of those which allows '/p/' to work, but not sure if that's correct. | 23:56 |
| *** tmcpeak has joined #openstack-infra | 23:56 | |
| jeblair | marun: i'm inclined to think (without having looked too deeply), that we should probably consider that a mis-feature and deprecate it. | 23:56 |
| *** timcline has joined #openstack-infra | 23:57 | |
| *** ddieterly has quit IRC | 23:57 | |
| *** ghostpl_ has quit IRC | 23:57 | |
| clarkb | zaro that confirms its an apache problem, the fix is what asselin did | 23:57 |
| marun | jeblair: I'm a bit confused by zuul's various job aggregation mechanisms. | 23:57 |
| marun | jeblair: I've noticed that skip-if matchers don't match on jobs that are defined by template, for instance. | 23:58 |
| zaro | clarkb: ahh, ok. i'll diff with asselin's change to see. | 23:58 |
| marun | jeblair: or maybe I'm just confused, but jobs are properly skipped in the check pipeline but then not skipped in the gate pipeline. | 23:58 |
| marun | jeblair: anyway, don't mean to distract if you're busy. I'll try to have more for tomorrow if you have some free time. | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
