| mordred | I put TONS of copied boilerplate stuff into each setup.py | 00:00 |
|---|---|---|
| mordred | we did this approach for a while - it sucked | 00:00 |
| *** danwent has quit IRC | 00:01 | |
| *** matiu has joined #openstack-dev | 00:01 | |
| *** matiu has quit IRC | 00:01 | |
| *** matiu has joined #openstack-dev | 00:01 | |
| *** noslzzp has quit IRC | 00:01 | |
| dstufft | mordred: If you really want d2to1 I don't really care. It's a dead project but it'll more or less work, d2to1 depends on setuptools which is wrong, but the machinary inside pbr is backwards and is going to cause pain (and this machinary existed prior to PBR and the same problems) | 00:02 |
| dstufft | or rather d2 is the dead project | 00:02 |
| mordred | dstufft: sure. I know d2 is a dead project | 00:02 |
| dstufft | no idea if d2to1 is still kicking or not | 00:02 |
| *** gyee has quit IRC | 00:03 | |
| mordred | d2to1 is still alive and kicking | 00:03 |
| mordred | if it dies, I'll just subsume it | 00:03 |
| mordred | however, | 00:03 |
| mordred | I'm willing to do something else | 00:03 |
| dstufft | There will probably be pain associated with moving from d2to1 to PEP426 | 00:03 |
| dstufft | once that's ready anyways | 00:03 |
| mordred | but a) PEP426 isn't ready | 00:03 |
| mordred | b) I need conversations about the subject to not start from the assumption that what I want to accomplish is incorrect | 00:03 |
| *** noslzzp has joined #openstack-dev | 00:03 | |
| mordred | because I've got eleventy-billion use cases and specific problems I'm trying to solve. if there is a better way tot solve them, AWESOME | 00:04 |
| mordred | honestly | 00:04 |
| mordred | not wed to an approach | 00:04 |
| mordred | taking off - back in a few | 00:06 |
| dstufft | mordred: k | 00:06 |
| *** sride has joined #openstack-dev | 00:06 | |
| *** alop has joined #openstack-dev | 00:07 | |
| smoser | vishy, do you have a cirros in a static route env that you could test something on? | 00:08 |
| *** esheffield has quit IRC | 00:08 | |
| smoser | edit /usr/share/udhcpc/default.script ; and then add at top 'set > /tmp/$1-info.txt' | 00:09 |
| smoser | i *think* you should have something in 'staticroutes' that we can then do something with. | 00:09 |
| *** andrewbogott is now known as andrewbogott_afk | 00:10 | |
| *** stevemar has quit IRC | 00:11 | |
| *** andrewbogott_afk is now known as andrewbogott | 00:12 | |
| *** alop has quit IRC | 00:12 | |
| *** andrewbogott is now known as andrewbogott_afk | 00:12 | |
| *** alop has joined #openstack-dev | 00:12 | |
| *** yidclare has quit IRC | 00:16 | |
| *** diogogmt has quit IRC | 00:18 | |
| smoser | vishy, if you can do that and open a bug, i dont think it'd be too hard. the stuff isn't documnted very well. | 00:19 |
| mordred | dstufft: k. I'm back. so - for starters, I'm super curious about the mechanism you refered to for a requirements.txt file to consume the setup.py information | 00:21 |
| ayoung | harlowja, did switching the backend in mysql make a differences? | 00:22 |
| harlowja | ayoung yup | 00:22 |
| harlowja | RH comes with myisam by default, and i forgot about that | 00:22 |
| *** reed has quit IRC | 00:22 | |
| harlowja | not sure what the intended effect should be though, myisam usage will die i think now ;) | 00:22 |
| ayoung | Ugh | 00:22 |
| harlowja | *or that migration will die on it* | 00:22 |
| ayoung | and we claim to be enterprise | 00:22 |
| harlowja | :) | 00:23 |
| harlowja | oh ya, u work for RH right ? | 00:23 |
| harlowja | oops | 00:23 |
| ayoung | I guess in the migrations we need to test if the constraint is there before dropping | 00:23 |
| * harlowja can never remember who people work for, haha | 00:23 | |
| mordred | we have stuff in nova to ensure innodb usage. perhaps we should make sure everything uses it? | 00:23 |
| dstufft | mordred: do you want an editable install or a normal install of whatever library is in setup.py:setup(name="Foo") | 00:23 |
| ayoung | Yeah, I wear the red hat | 00:23 |
| *** virajh has quit IRC | 00:23 | |
| ayoung | mordred, probably too late.... | 00:24 |
| *** epim has quit IRC | 00:24 | |
| mordred | dstufft: normal. | 00:24 |
| dstufft | mordred: . | 00:24 |
| *** epim_ has joined #openstack-dev | 00:24 | |
| dstufft | (literally just a period) | 00:24 |
| mordred | hahahaahaha | 00:24 |
| mordred | yes. ok | 00:24 |
| mordred | :) | 00:24 |
| mordred | so - what good does that do me? | 00:24 |
| dstufft | if you want editable it's -e . | 00:25 |
| *** roampune has quit IRC | 00:25 | |
| *** colinmcnamara has quit IRC | 00:25 | |
| mordred | I mean, I can see how that would let me have a file called requirements.txt that would cause setup.py to install the depends | 00:25 |
| mordred | but nothing consumes that file | 00:26 |
| harlowja | ayoung i can help test it again if u need | 00:26 |
| mordred | all of the tooling entry points are via setup.py commands? | 00:26 |
| openstackgerrit | A change was merged to openstack/glance: python3: Introduce py33 to tox.ini https://review.openstack.org/31396 | 00:27 |
| openstackgerrit | A change was merged to openstack/python-glanceclient: python3: Introduce py33 to tox.ini https://review.openstack.org/31413 | 00:27 |
| dstufft | mordred: The typical workflow is for something you expect to publish via PyPI is that you declare your deps in setup.py, if someone clones that and wants to install it then can just do pip install -r requirements.txt, but all the tooling in the python world operates off the setup.py. The processing of requirements.txt just adds indrection | 00:27 |
| dstufft | mordred: you also no longer need to handle parsing a requirements.txt, which will remove some code from pbr | 00:27 |
| dstufft | and you let pip handle processing it | 00:27 |
| dstufft | it also makes it super simple to handle development overrides | 00:27 |
| mordred | dstufft: I totally grok the typical workflow - we started off that way | 00:28 |
| mordred | the nice thing about having a reqiurements file that's not in setup.py | 00:29 |
| dstufft | because a setup.py only references a name, that name can come from anywhere, PyPI, github, whatever. so a common idiom is to say depend on Django in install_requires, and then in your requirements.txt you can specify an exact index to install it from. | 00:29 |
| *** danwent has joined #openstack-dev | 00:29 | |
| dstufft | mordred: well you can use d2to1 to put it in setup.cff if you really want | 00:29 |
| mordred | :) | 00:29 |
| mordred | in general, we don't realy make use of the other-index features | 00:30 |
| ayoung | harlowja, ok...although I suspect I can find an RHEL machine around here somewhere | 00:30 |
| harlowja | ;) | 00:30 |
| mordred | except environmentally, in that we run a pypi mirror in our CI because pypi can't deal with us | 00:30 |
| mordred | but we don't encode that in the actual trees themselves | 00:30 |
| mordred | because pypi is fine for normal installation | 00:31 |
| dstufft | mordred: you don't, but other people may. Maybe I fork python-keystone client and want to point to a different index, or I want to override a particular dependency, if both of us are using requirements.txt then it becomes easy to confuse what I added and what comes from upstream | 00:31 |
| dstufft | mordred: I'd be interested to know what "pypi can't deal with us" means, since i'm a PyPI admin and would love to fix it | 00:31 |
| mordred | eternal links | 00:32 |
| dstufft | they are being killed | 00:32 |
| dstufft | :) | 00:32 |
| mordred | Iknow | 00:32 |
| mordred | thkan god | 00:32 |
| mordred | :) | 00:32 |
| dstufft | with much wailing and gnashing of teeth | 00:32 |
| clarkb | there are other problems too | 00:32 |
| mordred | we build MANY MANY MANY times a day. external links invaraible lead to build failures | 00:32 |
| dstufft | and more then a few angry emails | 00:32 |
| mordred | yah | 00:32 |
| mordred | thank you for your work on that | 00:32 |
| clarkb | pypi itself has had not so great uptime | 00:32 |
| dstufft | clarkb: It does now | 00:32 |
| mordred | I mean, we'll probably still run a mirror | 00:32 |
| dstufft | :) | 00:32 |
| clarkb | and other people have a habit of uploading crap | 00:32 |
| mordred | just because of our build load | 00:33 |
| *** comay has joined #openstack-dev | 00:33 | |
| dstufft | I mean I don't mind that you run a mirror, I just want it to be a choice and not something you felt forced to do | 00:33 |
| mordred | totally | 00:33 |
| clarkb | dstufft: /me sidetracks the conversation. it would be super duper if pypi uploads didn't feel so hacky. Does https work for that yet? | 00:34 |
| dstufft | mordred: I'm looking at pbr and it looks like 1/3 or 1/2 of the code here exists to deal with parsing a requirements.txt into a setup.py, so that's a bunch of code you could delete immediately switching to declaring in setup.py|cfg | 00:34 |
| mordred | dstathat's actually a pretty small amount of the code | 00:34 |
| dstufft | clarkb: It *works", but distutils doesn't verify it (or use it until 3.4) | 00:34 |
| dstufft | Not sure if setuptools added verification of that or not | 00:34 |
| mordred | dstufft: we upload manually | 00:35 |
| clarkb | dstufft: well we stopped using distutils for it :) | 00:35 |
| dstufft | I have on my stack to write a seperate utility to upload | 00:35 |
| clarkb | dstufft: currently we use curl | 00:35 |
| clarkb | I should update the url to try https again | 00:35 |
| clarkb | it wasn't working when we wrote that script | 00:35 |
| dstufft | and then convince and/or threaten and/or do it without asking to break the existing API since almost everything uses either HTTP or unverified HTTPS | 00:35 |
| dstufft | clarkb: if it doesn't work let me know, I'm always on freenode IRC as dstufft or donald@dstufft.io // donald@python.org | 00:36 |
| dstufft | donald@stufft.io* | 00:36 |
| clarkb | dstufft: will do, thanks | 00:36 |
| mordred | dstufft: we do other things- we set versions, we set sphinx defaults, we do package finding, we generate authors and changelog files | 00:36 |
| *** sleepsonthefloor has quit IRC | 00:36 | |
| mordred | dstufft: I should honestly write up a "here is what we want to accomplish" doc to point you at | 00:36 |
| *** wenjianhn has joined #openstack-dev | 00:36 | |
| mordred | dstufft: because what I have now is a description of what we do | 00:36 |
| mordred | not why | 00:36 |
| dstufft | mordred: sure | 00:37 |
| dstufft | mordred: You tell me what you want to accomplish and I can work from there. | 00:37 |
| mordred | but most all of it has come from writing code to make automation and dev workflow work better | 00:37 |
| mordred | then abstracting that out | 00:37 |
| mordred | but seriously - I'm SUPER happy with d2to1 | 00:38 |
| mordred | implementaiton not withstanding | 00:38 |
| dstufft | I'm not as opposed to d2to1 as I am about some of the other parts | 00:38 |
| mordred | but the interface that it gives a project | 00:38 |
| *** changbl has joined #openstack-dev | 00:38 | |
| mordred | and operating a project that uses it | 00:38 |
| dstufft | I don't like it, but it's not in and of itself a bad idea | 00:38 |
| dstufft | mordred: but yea, shoot me a doc of what you're trying to achieve (vs what you're doing) and I'll see what sort of proposal I can come up with, either of those addresses or donald.stufft@nebula.com works too if you want a more officialish email | 00:40 |
| mordred | dstufft: ah! I didn't realize you were at nebula | 00:40 |
| dstufft | mordred: ah yea, since Oct '12 | 00:41 |
| *** Q310 has quit IRC | 00:42 | |
| dstufft | also more or less any pain points that are had with packaging or PyPI you can send my way. I said earlier not sure if you saw. I'm more or less one of the primary folks working on this on the python side, so I'm as good as person as any to contact | 00:42 |
| mordred | dstufft: well, good to know | 00:43 |
| dstufft | also I have a high pain tolerance so if I believe in something I smash my face against it until either my face or the problem breaks | 00:43 |
| mordred | dstufft: I'm sure you will not like the patches currently in queue for pbr :) | 00:43 |
| *** Qten has joined #openstack-dev | 00:43 | |
| clarkb | what about pip? | 00:43 |
| dstufft | clarkb: sure pip too, I'm core dev on pip | 00:43 |
| clarkb | because it would be really cool if pip didn't isntall things in a depth first search | 00:43 |
| *** darjeeling has joined #openstack-dev | 00:44 | |
| dstufft | There might be a reason for that, I'd have to look | 00:44 |
| dstufft | You guys will probably like this PR https://github.com/pypa/pip/pull/985 | 00:44 |
| mordred | dstufft: yess | 00:45 |
| clarkb | and if it could somehow monkey patch itself in place of easy_install so that transitive deps don't derp (not sure how possible that is) | 00:45 |
| *** darjeeling has quit IRC | 00:45 | |
| mordred | dstufft: I'm also awaiting a released version with wheel support | 00:45 |
| *** darjeeling has joined #openstack-dev | 00:45 | |
| dstufft | mordred: yea that'll be 1.4 too | 00:45 |
| mordred | dstufft: yup | 00:45 |
| dstufft | along with ^ if I get my way | 00:45 |
| mordred | has my vote | 00:45 |
| *** jecarey has quit IRC | 00:45 | |
| mordred | clarkb: well, the latest pbr patches remove easy_install from the equation for our stuff :) | 00:45 |
| dstufft | clarkb: Can you file a ticket about DFS and why? (or if there is one add my name to it with a mention) | 00:45 |
| dstufft | mordred: oh thank god | 00:46 |
| dstufft | That bit of code made me majorly sad | 00:46 |
| *** jecarey has joined #openstack-dev | 00:46 | |
| clarkb | dstufft: I can but it isn't complicated. DFS fails when it installs foo then tries to isntall foo <2.0. | 00:46 |
| mordred | pip makes me quite happy, and easy_install makes me want to die inside | 00:46 |
| clarkb | if the latest version of foo is >=2.0 | 00:46 |
| dstufft | I don't think easy_install does SSL verification yet | 00:46 |
| *** stevemar has joined #openstack-dev | 00:47 | |
| mordred | dstufft: actually, although I'm sure the intent of the latest set of pbr patches might make you die inside - but I wouldn't mind eyeballs on any obvious "that's actually going to physically break things" sorts of problems | 00:47 |
| dstufft | mordred: link? | 00:47 |
| mordred | dstufft: https://review.openstack.org/#/c/31783/ and https://review.openstack.org/#/c/30521/ and https://review.openstack.org/#/c/32222/ and https://review.openstack.org/#/c/30522/ | 00:48 |
| mordred | dstufft: ignore the jenkins failures - I'll fix that of course :) | 00:48 |
| dstufft | mordred: is there a way to make gerrit show me all the diffs at once in one page | 00:49 |
| mordred | dstufft: no. it's on the todo list | 00:49 |
| dstufft | Diff All gave me like 6 windows | 00:49 |
| dstufft | ah ok | 00:49 |
| mordred | dstufft: nokia wrote the patch and then wouldn't sign the google cla to contriute it :) | 00:50 |
| dstufft | distribute<0.7 got pulled IIRC | 00:50 |
| dstufft | fwiw | 00:50 |
| * mordred cries | 00:50 | |
| dstufft | (also a dead project) | 00:50 |
| dstufft | er | 00:50 |
| mordred | yah. I just put that there for safety | 00:50 |
| dstufft | I mean distribute==0.7 | 00:50 |
| dstufft | sorry | 00:50 |
| mordred | I'll do setuptools .7 once it's good | 00:50 |
| mordred | but last I heard it was having flaky upgrade problems this week | 00:51 |
| dstufft | iirc upgrading setuptools works, but distribute doesn't | 00:51 |
| dstufft | https://github.com/pypa/pip/issues/986 | 00:51 |
| *** seanrob has quit IRC | 00:52 | |
| mordred | but upgrading setuptols _from_ distribute? | 00:52 |
| dstufft | broken | 00:53 |
| dstufft | only setuptools -> setuptools works | 00:53 |
| dstufft | I believe | 00:53 |
| mordred | k. yeah. have to wait til that's fixed | 00:53 |
| mordred | way too much existing distribute install base for us right now | 00:54 |
| dstufft | mordred: Ok the first one makes me sad | 00:54 |
| dstufft | Do you only support enviornments with pip installed? | 00:54 |
| mordred | yes | 00:54 |
| mordred | I don't care about ancient python | 00:54 |
| dstufft | No python comes with pip ;) | 00:54 |
| dstufft | not until (hopefully) 3.4 | 00:54 |
| mordred | well, poorly stated on my partperhpas, but since easy_install is an absolute turd | 00:55 |
| dstufft | but as long as a requurement on pip is ok with you I think it will work (for some defintiion of work) | 00:55 |
| dstufft | also I suck at spelling | 00:55 |
| mordred | cool | 00:55 |
| dstufft | sorry | 00:55 |
| mordred | spelling sux | 00:55 |
| *** zodiak has joined #openstack-dev | 00:55 | |
| mordred | dstufft: python: No module named pip.__main__; 'pip' is a package and cannot be directly executed <-- did I derp? or did pip change a version that I missed? | 00:55 |
| mordred | oh - shit. I have pip 1.4 dev installed locally | 00:56 |
| dstufft | oh right | 00:56 |
| dstufft | that's a new thing in 1.4 | 00:56 |
| mordred | CRAP | 00:56 |
| * mordred cries | 00:56 | |
| dstufft | I've been using 1.4 for so long | 00:56 |
| mordred | ok. well, I know where to look at least | 00:56 |
| clarkb | mordred: yeah you probably won't get that fixed | 00:56 |
| mordred | it's fine - I can work around it | 00:56 |
| mordred | hrm. no. there's a __main__ in pip 1.3 | 00:57 |
| dstufft | mordred: I also object to the word sensible in these patches ;P | 00:57 |
| mordred | dstufft: I think we'll need to have that argument over beer | 00:58 |
| *** jecarey has quit IRC | 00:58 | |
| mordred | dstufft:and we really should - because I'm honestly bogled that you like the other approach | 00:58 |
| clarkb | dstufft: https://github.com/pypa/pip/issues/988 | 00:59 |
| dstufft | clarkb: thanks | 00:59 |
| dstufft | clarkb: I'm gonna add it to the 1.5 milestone, it might get knocked off there but that is a pretty big wart | 00:59 |
| dstufft | mordred: sure :) I'm pretty bad at explaining things and maybe I don't full grok the intentions here. But it feels like you're going further down the road of reimplementing parts of pip inside of pbr :) | 01:00 |
| mordred | dstufft: well, I'm more likely to reimplement parts of setuptools inside of pbr - and I wouldn't mind a complete clean re-impl | 01:00 |
| mordred | but | 01:00 |
| mordred | I don't want to duplicate pip if I can help it | 01:00 |
| clarkb | dstufft: https://github.com/pypa/pip/issues/989 is more minor but would be nice to fix | 01:00 |
| mordred | becuase I like pip | 01:01 |
| mordred | dstufft: but that's why I'm thinking beer - because if we can grok each other's intent - then working on a solution will make WAY more sense | 01:01 |
| mordred | :) | 01:01 |
| lifeless | mmm beer | 01:02 |
| dstufft | mordred: fwiw I wouldn't bother trying to do a clean room impl of setuptools. It's destined for being able to be replaced | 01:02 |
| dstufft | not sure how closely you followed the plans there | 01:02 |
| mordred | dstufft: I have followed the plans some, then I gave up | 01:02 |
| dstufft | PEP426 is installation (not developer side) metadata, so ``pip install`` would use it, developers don't edit it | 01:03 |
| dstufft | and then setuptools will be updated to generate it when you do setup.py sdist, bdist etc | 01:03 |
| dstufft | so installation no longer requires executing code to fetch metadata | 01:03 |
| mordred | ++ | 01:03 |
| mordred | I am 100% in favor of this | 01:03 |
| dstufft | After that, well any tool that generates the target format with a pymeta.json file can be used on the developer side | 01:04 |
| dstufft | it's a bit more complicated with sdist vs bdist | 01:04 |
| mordred | yah. so I imagine that I'll move pbr towards being able to do that then - but I'll try to keep up with what 426 is doing | 01:04 |
| dstufft | but that's the general dist | 01:04 |
| mordred | sure | 01:04 |
| dstufft | gist | 01:04 |
| * mordred no care about bdist, thank heavens | 01:04 | |
| dstufft | well wheel is a bdist | 01:04 |
| dstufft | there's a bunch of different ones | 01:04 |
| mordred | ok. me care about bdist | 01:05 |
| mordred | via wheel | 01:05 |
| mordred | but I'm also happy to use pip to generate wheels for me | 01:05 |
| mordred | at least for a while | 01:05 |
| dstufft | Sad thing is wheels are less useful for linux | 01:05 |
| mordred | the're TOTALLY useful | 01:05 |
| mordred | I cannot WAIT | 01:05 |
| mordred | they are going to be a huge win for our build farm | 01:05 |
| dstufft | well I mean from the developer -> end user flow | 01:06 |
| *** kaushikc has joined #openstack-dev | 01:06 | |
| mordred | you know what's stupid? recopiling lxml 2000+ times a day | 01:06 |
| dstufft | they are huge for endusers, but you can't really publish a linux wheel on pypi | 01:06 |
| mordred | no | 01:06 |
| mordred | certainly not | 01:06 |
| dstufft | Although I have ketches for a wheel build farm that has Os specific end points | 01:06 |
| dstufft | sketches | 01:06 |
| *** devoid has quit IRC | 01:06 | |
| dstufft | if I ever get around to doing it | 01:06 |
| mordred | we should talk | 01:06 |
| mordred | we run a REALLY big build famr here you know | 01:07 |
| dstufft | (I also run Crate.io :V) | 01:07 |
| mordred | ah - I looked at using that, btw | 01:07 |
| dstufft | it sucks, it just sucked less than PyPI | 01:07 |
| lifeless | dstufft: why can't we publish a linux wheel on pypi ? | 01:07 |
| dstufft | lifeless: because a wheel built on ubuntu won't nesc run on RHEL | 01:07 |
| mordred | lifeless: soname | 01:07 |
| mordred | lifeless: you'd have to do them per shared lib combo | 01:07 |
| lifeless | dstufft: that sounds like a wheel bug | 01:07 |
| dstufft | because binary compatibility isn't a thing on Linux with shared libs (or something) | 01:08 |
| dstufft | I don't know C | 01:08 |
| dstufft | so i'm regurgitating | 01:08 |
| mordred | dstufft: btw - the wheel hook to allow for executing code on wheel install is a bug | 01:08 |
| dstufft | mordred: an installation hook is a hard requirement for some projects | 01:08 |
| *** esp has joined #openstack-dev | 01:08 | |
| lifeless | dstufft: ok, so sure they need to be platform specific, but 'rhel 5 i386' is no worse than 'windows xp i386' | 01:08 |
| mordred | dstufft: I disagree - I think an install hook is a workaround for current brokeness | 01:09 |
| *** cloudnull has left #openstack-dev | 01:09 | |
| lifeless | dstufft: so I don't understand why windows, or Mac OSX would be ok, but linux wouldn't be. | 01:09 |
| mordred | dstufft: and I think that fixing that brokenness for real would be more beneficial | 01:09 |
| *** kaushikc1 has joined #openstack-dev | 01:09 | |
| lifeless | mordred: plugin registration caches is the thing, for twisted. Not saying I agree; just that that is the scenario. | 01:09 |
| mordred | dstufft: but now we're ratholing | 01:09 |
| dstufft | mordred: Glyph (of Twisted Fame) was very vocal about needing one | 01:09 |
| mordred | lifeless: I know | 01:09 |
| *** kaushikc has quit IRC | 01:09 | |
| mordred | but that's because plugin registration caches are broken | 01:09 |
| mordred | let's fix them | 01:09 |
| lifeless | dstufft: to be fair though, the twisted plugin system is super slow. | 01:09 |
| dstufft | mordred: but I plan on adding a flag to pip (sooner or later) that lets you say "don't run hooks" | 01:10 |
| lifeless | dstufft: so I think its self inflicted :) | 01:10 |
| mordred | dstufft: AWESOME. thank you | 01:10 |
| *** terryg__ has quit IRC | 01:10 | |
| mordred | lifeless: ++ | 01:10 |
| dstufft | lifeless: I really don't know enough about C or binary compatability to speak to it, the relevant PEP is http://www.python.org/dev/peps/pep-0425/ | 01:10 |
| *** jamielennox|away is now known as jamielennox | 01:11 | |
| dstufft | I think it was designed mostly to signal compatability with a particular Python | 01:11 |
| dstufft | ABI or something | 01:11 |
| *** benner has quit IRC | 01:11 | |
| dstufft | *says words he only half understands* | 01:11 |
| mordred | dstufft: is there any chance you'd consider taking a patch to pip that would add support for markerlib stanzas in requirements files? (I'm guessing no given the preceeding conversation) | 01:11 |
| clarkb | going back to requirements.txt. why would you want that to determine where things are fetched from but not actually list the requirements. It seems like a super poorly named file if that is the case | 01:11 |
| clarkb | my pip.conf tells pip where to grab things from, requirments.txt lists the requirments of my project | 01:12 |
| lifeless | dstufft: righto, thats totally able to create a cython module wheel for pypi as I read it. | 01:12 |
| lifeless | dstufft: its external libs that will cause headaches. | 01:12 |
| *** benner has joined #openstack-dev | 01:12 | |
| *** yjiang5 has quit IRC | 01:12 | |
| dstufft | clarkb: I'm really bad at explaining this, http://yehudakatz.com/2010/12/16/clarifying-the-roles-of-the-gemspec-and-gemfile/ does a better job, just pretend Gemfile is requirements.txt and gemspec is setup.py | 01:13 |
| dstufft | nd pretend gemfile.lock doesn't exist for reasons | 01:13 |
| mordred | dstufft: I tried reading that | 01:13 |
| mordred | and quite honestly, not trying to be a jerk | 01:13 |
| mordred | but it was greek | 01:13 |
| dstufft | heh | 01:13 |
| mordred | because I do not know anything about ruby | 01:14 |
| mordred | or how gemfiles or gemspecs work - except that they fail all the time for me | 01:14 |
| mordred | so, basically, I consider them a broken system | 01:14 |
| mordred | clarkb: if the above makes sense to you, can you translate for me?) | 01:14 |
| dstufft | So paraphrasing, more or less requirements.txt are for "applications", it so happens that when you're developing a library the library _is_ the application | 01:14 |
| clarkb | mordred: I think basically ruby confuses packages and package repos | 01:15 |
| clarkb | and puts package repo management in package repos | 01:15 |
| lifeless | clarkb: ruby confuses. | 01:15 |
| *** bknudson has quit IRC | 01:15 | |
| clarkb | er gets around repo management by enforcing that sort of thing local to the package | 01:15 |
| dstufft | and when you depend on something for an application, you what a very specific set of requirements | 01:15 |
| clarkb | which seems insane | 01:15 |
| *** rcleere has joined #openstack-dev | 01:15 | |
| dstufft | e.g. it's common for a requirements.txt to pin to exact versions | 01:15 |
| dstufft | a requirements.txt also pairs a name (requests==1.0) to a location (pypi.python.org) | 01:16 |
| clarkb | "While the application author may not have cared much (up front), which version of nokogiri the application used, he does care a lot that the version used in development will remain the version used in production." is the important bit I think | 01:16 |
| clarkb | and I don't think a package should care | 01:16 |
| dstufft | But a setup.py should have very widely defined dependencies, as wide as possible | 01:16 |
| clarkb | but maybe I spend too much time with apt | 01:16 |
| mordred | dstufft: that's where the disagreement is | 01:16 |
| mordred | dstufft: because I agree with the theory that you say | 01:16 |
| mordred | but in practice, it doesn't work | 01:17 |
| mordred | because broad depends break ALL the TIME in python | 01:17 |
| mordred | because of how people release crap to pypi | 01:17 |
| mordred | so I want my setup.py to be prescriptive, so that the thing you install actually works | 01:17 |
| mordred | because I tested that it works | 01:17 |
| dstufft | mordred: well then you get into dependency hell | 01:17 |
| dstufft | which is mitigated with venvs | 01:18 |
| mordred | dstufft: not really | 01:18 |
| mordred | oh - well, yeah. lemme rephrase | 01:18 |
| dstufft | mordred: foo depends on requests>=1.0, now I want to install bar and it depends on requests<1.0 | 01:18 |
| mordred | right. but that's a python packaging failing | 01:18 |
| *** kaushikc1 has quit IRC | 01:18 | |
| clarkb | dstufft: assuming those values are correct they won't work with your broad declarations | 01:19 |
| mordred | doesn't help if you dont' declare the versioned depends | 01:19 |
| clarkb | yeah | 01:19 |
| dstufft | bad example-- | 01:19 |
| mordred | if foo needs >= 1.0 of requests because of API | 01:19 |
| dstufft | foo depends on requests==1.1.1 | 01:19 |
| dstufft | bar depends on requests==0.9.0 | 01:19 |
| dstufft | or bar depends on requests==1.1.0 | 01:19 |
| *** novas0x2a|laptop has quit IRC | 01:19 | |
| mordred | right. so - a think should depend on as broad a range as it can actually work with | 01:19 |
| *** jasdeepH has quit IRC | 01:20 | |
| dstufft | sure, which is what I said :) as wide as possible, but not so wide it doesn't work :) | 01:20 |
| mordred | but that range, that it declares, should be the thing that it does work with, so that's what the install mechanisms should use to install it | 01:20 |
| mordred | yes | 01:20 |
| mordred | we agree on that | 01:20 |
| dstufft | a requirements.txt file you generally want to pin to exact versions | 01:20 |
| mordred | yes | 01:20 |
| mordred | no | 01:20 |
| mordred | why? | 01:20 |
| clarkb | mordred: so that you don't need a package repo | 01:20 |
| *** tjones has left #openstack-dev | 01:20 | |
| *** danwent has quit IRC | 01:20 | |
| mordred | I want the thing we just discussed. as wide as possible, but not so wide it doesn't work :) | 01:20 |
| mordred | why would I want something more prescriptive in requirements.txt | 01:21 |
| mordred | ? | 01:21 |
| clarkb | dstufft: and gems are claiming you want exact versions in requirements.txt so that when you install that in your dev env and in your prod env they work the same | 01:21 |
| clarkb | this is a problem solved by package repos not by packages | 01:21 |
| mordred | AH | 01:21 |
| mordred | yes | 01:21 |
| mordred | they are trying to solve a problem for small scale projects | 01:21 |
| dstufft | requirements.txt doesn't belong inside a package at all ;P | 01:21 |
| mordred | we are... a bit larger than that :) | 01:21 |
| mordred | dstufft: well not, not by your description of its use | 01:22 |
| mordred | totally agree | 01:22 |
| dstufft | A package repo can have multiple versions of a dependency | 01:22 |
| dstufft | not sure how it solves it | 01:22 |
| mordred | you test against the repository | 01:22 |
| mordred | your dev env gets the version that's in the repository | 01:22 |
| mordred | you don't need a specific versoin in the requirements.txt file | 01:23 |
| mordred | unless what you are saying | 01:23 |
| mordred | is taht you dev with the broad reqs | 01:23 |
| dstufft | until someone upgrades a dependency in the repo for some other reason | 01:23 |
| mordred | and then generate the requirements.txt file | 01:23 |
| harlowja | oh, this looks like a fun conversation | 01:23 |
| dstufft | unless you have a seperate repo for every project | 01:23 |
| mordred | and then use that file to install your prod env | 01:23 |
| mordred | is that the idea? | 01:23 |
| mordred | so that you make sure your prod install installs the versoins you developed with | 01:24 |
| dstufft | mordred: more or less, you don't generally care exactly what version you get when you install it in the dev environment, you do care exactly (Down the minor-est version number) that prod matches dev | 01:24 |
| mordred | that is a workflow that makes sense for a certain class of things, for sure | 01:24 |
| dstufft | you might care within a range on dev | 01:24 |
| mordred | right | 01:24 |
| *** melwitt has quit IRC | 01:24 | |
| mordred | but we have 800 devs | 01:24 |
| *** pixelbeat has joined #openstack-dev | 01:24 | |
| mordred | which means that there is no place from which togenerate that file | 01:25 |
| mordred | it's a workflow that doesn't match us | 01:25 |
| dstufft | mordred: the requirements.txt would be for the people installing OpenStack inside of a company | 01:25 |
| mordred | we need to describe the set of requirements that the particular project needs | 01:25 |
| dstufft | it's a deployment file | 01:25 |
| mordred | dstufft: they can generate one if they want | 01:25 |
| harlowja | dstufft i've done the following in anvil, and its what yahoo uses currently for installs, anvil checks out all openstack components it wants code, anvil examines there 'requires' (without installing) and reports if the dependencies just won't work together, then anvil looks up in the repositories for packages that will satisfy those requires if they exist there (version checking and all) and if missing ones are found it builds | 01:25 |
| harlowja | those, then it creates a single repository of all dependencies + the openstack components, that becomes sorta what u are talking about with a single repo of everything (aka the deployment) | 01:25 |
| mordred | dstufft: we're still talking about two different things | 01:26 |
| harlowja | i think others do something similar with a deployment to a venv and ship that | 01:26 |
| mordred | and I think it might be hinging on the fact that we're calling the file requirements.txt | 01:26 |
| mordred | which, to you, is a file that should get generated | 01:26 |
| mordred | go back 1 month when this file was called tools/pip-requires | 01:27 |
| mordred | does it bother you in the same way then? | 01:27 |
| *** epim_ has quit IRC | 01:27 | |
| dstufft | mordred: that's what a python developer thinks of when they see the requirements.txt file :) If you had the contents of that file directly inside of setup.py|cfg then sure whatever. If you were pinning to exact versions i'd probably argue against that (not sure if you were?) but requirements.txt is a pip file that has a particular purpose | 01:27 |
| harlowja | mordred doesn't the requiresments.txt eventually get shipped in the egg_info of the final package anyway? | 01:27 |
| dstufft | harlowja: it does for openstack, but that's backwards for how Python packages generally work | 01:28 |
| harlowja | ah | 01:28 |
| mordred | dstufft: so, I'd beg to differ | 01:28 |
| lifeless | dstufft: how do folk normally obtain the file then ? | 01:28 |
| harlowja | didn't know it was backwards, i guess they usually let the repo package specify the dependencies? | 01:28 |
| lifeless | dstufft: cause, I've only ever seen the file at all in openstack. | 01:28 |
| harlowja | *aka rpm/deb? | 01:28 |
| dstufft | harlowja: normally the requirements.txt is a per project file, not a library file | 01:29 |
| mordred | in that I'm seeing more and more people putting requiremets.txt into their repos | 01:29 |
| *** mrodden has joined #openstack-dev | 01:29 | |
| dstufft | mordred: normally those are developer oriented for when working _on_ the package, not for installing the package as a dep of something | 01:29 |
| dstufft | e.g. I do that | 01:29 |
| *** Mandell has quit IRC | 01:29 | |
| *** jbresnah has joined #openstack-dev | 01:30 | |
| dstufft | https://gist.github.com/dstufft/5753926 one my requirements.txt files from a library | 01:30 |
| harlowja | interesting, wish there was some kind of python standard about requirements :( | 01:30 |
| *** jecarey has joined #openstack-dev | 01:30 | |
| dstufft | lifeless: normally you won't see it becasue it's specific to a particular deployment | 01:30 |
| dstufft | (generally) | 01:30 |
| dstufft | openstack is the oddball there :) | 01:30 |
| lifeless | dstufft: there's an analagous file in buildout environments, which they usually ship with the app | 01:31 |
| mordred | dstufft: right. what I'm saying is that I do not think the "standard" way" people do this scales for large dev | 01:31 |
| mordred | I could be VERYU wrong of course | 01:31 |
| clarkb | mordred: how can it if everyone is using a different requirements file | 01:31 |
| dstufft | mordred: I don't see how putting your deps in setup.py|cfg is any harder than putting them in requirements.txt | 01:31 |
| lifeless | dstufft: how would folk putting those apps on pypi do it? | 01:31 |
| clarkb | now foo fails for bar and baz can't reproduce | 01:31 |
| mordred | dstufft: well, case in point..., | 01:32 |
| harlowja | i wonder if the thing u are talking about is why its currently hard for me to do rpmbuild correctly, that command executes python setup.py install --root=X internally, and then when that runs it tries to pull in all the depenendencies (this seems unique to openstack?) and that kills the rpmbbuild (since it complains about files not packaged, aka the dependencies the install downloaded) | 01:32 |
| harlowja | *but maybe not* | 01:33 |
| mordred | harlowja: no. we're talking about semantics around where we put the list of things that gets sucked in | 01:33 |
| harlowja | k, haha | 01:33 |
| dstufft | mordred: you're essentially already listing them in setup.py, but you're piggybacking off a file that has an generally accepted purpose. It becomes confusing to people not familar with Openstack :) | 01:33 |
| mordred | dstufft: so, the generally accepted purpose is not documented, nor is it clear to folks | 01:33 |
| mordred | dstufft: I NEVER would have imagined the workflow you talk about | 01:33 |
| mordred | dstufft: and when I see one of those files, I assume it contains information I'd want to consume | 01:34 |
| clarkb | and there is a reason for splitting the requirements outside of setup.py | 01:34 |
| clarkb | because easy_install | 01:34 |
| mordred | yeah, also, easy_install sucks | 01:34 |
| harlowja | +1 | 01:35 |
| *** krtaylor has joined #openstack-dev | 01:35 | |
| mordred | I think we're back to the rathole - especially with me on a laggy connection | 01:35 |
| *** otherwiseguy has joined #openstack-dev | 01:35 | |
| dstufft | you're reading a requirements.txt _into_ setup.py install_requires, there is no functional difference for easy_install in this case. It just muddies the water. You could literally copy/paste the stuff from requirements.txt into setup.py|cfg, and add a "." in requirements.txt and the confusion goes away and no changes in functionality (plus you can delete support code meant to read a requirements.txt) | 01:36 |
| dstufft | mordred: THere's http://www.pip-installer.org/en/latest/requirements.html but it's not exactly the best docu | 01:36 |
| mordred | dstufft: there is a difference | 01:36 |
| mordred | dstufft: pip supports thigns in that file that easy_install does not | 01:36 |
| dstufft | mordred: what's the functional difference? | 01:36 |
| dstufft | mordred: if easy_install doesn't support it then setup.py doesn't support it and someone doing ``pip install yourpackage`` isn't going to get it, pip doesn't read requirements when you type ``pip install yourpackage`` | 01:37 |
| ayoung | harlowja, https://review.openstack.org/#/c/32510/ | 01:37 |
| *** pixelbeat has quit IRC | 01:37 | |
| mordred | but what you want is a file in pip format, with those features, that can be the basis of installation | 01:37 |
| harlowja | ayoung woot | 01:37 |
| mordred | dstufft: it does with my new patch ot pbr :) | 01:37 |
| dstufft | s/requirements/requurements.txt/ | 01:37 |
| clarkb | also, if requirements.txt is used for making things more restrictive what prevents you from doing that with the way we use requirements.txt. It seems like you could just edit it as you would normally to get that behavior | 01:37 |
| ayoung | harlowja, wait, something is missing from that commit | 01:38 |
| harlowja | cool, i just added a comment | 01:38 |
| harlowja | might be the same thing :-p | 01:38 |
| dstufft | mordred: well now you're not even using pip anymore and you're using pip-openstack | 01:39 |
| dstufft | if you're going to change the behavior of pip | 01:39 |
| mordred | dstufft: I'm using stock pip | 01:39 |
| dstufft | mordred: how are you going to get something that setup.py doesn't support, but pip does from a requirements.txt into pip via setup.py | 01:39 |
| *** erkules_ has joined #openstack-dev | 01:39 | |
| ayoung | harlowja, somehow returning None got dropped from get_fkey_constraint_name | 01:39 |
| ayoung | must have hit undo too many times | 01:39 |
| harlowja | :) | 01:40 |
| mordred | dstufft: because setup.py in my projects will not use easy_install to do anything | 01:40 |
| *** lonelyibex has joined #openstack-dev | 01:40 | |
| mordred | (we may be talking past each other) | 01:40 |
| dstufft | mordred: easy_install is not executed if you pip install something, at all | 01:40 |
| clarkb | dstufft: it is | 01:40 |
| dstufft | unless your setup.py does it | 01:40 |
| *** nati_ueno has quit IRC | 01:40 | |
| clarkb | dstufft: transitive dependencies | 01:40 |
| dstufft | setuptools != easy_install | 01:41 |
| clarkb | it uses the same mechanisms no? | 01:41 |
| clarkb | re pip supporting things the others do not | 01:41 |
| mordred | clarkb: no, pip does an egg_info command and then reads the requires.txt file in the egg-info dir | 01:41 |
| mordred | dstufft: right? ^^? | 01:41 |
| dstufft | yes | 01:41 |
| mordred | right. that' stotally sane | 01:41 |
| mordred | what's problematic is when someone grabs your repo or tarball by hand | 01:42 |
| mordred | and then does "python setup.py install" and that borks | 01:42 |
| *** erkules has quit IRC | 01:42 | |
| *** maoy has joined #openstack-dev | 01:42 | |
| clarkb | mordred: then why do we need to set our mirror in .pydistutils.cfg? | 01:43 |
| mordred | clarkb: setup_requires | 01:43 |
| dstufft | If you're using setup_requires setuptools does that when you do egg_info | 01:43 |
| dstufft | mordred: i'm still interested in how you're going to take something defined in requirements.txt and tell pip about it without modifying behavior :D | 01:44 |
| dstufft | that setup.py doesn't natively support* | 01:44 |
| openstackgerrit | A change was merged to openstack-infra/devstack-gate: Convert README from markdown to reStructuredText. https://review.openstack.org/32456 | 01:45 |
| openstackgerrit | A change was merged to openstack/keystone: Remove bufferedhttp https://review.openstack.org/31973 | 01:45 |
| openstackgerrit | A change was merged to openstack/swift: Restore loadapp checking https://review.openstack.org/31075 | 01:45 |
| *** erkules_ has quit IRC | 01:45 | |
| mordred | dstufft: ok. fair. we were talking past each other - I see what you're asking now | 01:46 |
| mordred | dstufft: you're on "what do you gain by putting the list in external files rather than as python lists in setup.py" yes? | 01:46 |
| *** prad has joined #openstack-dev | 01:47 | |
| *** erkules_ has joined #openstack-dev | 01:47 | |
| dstufft | mordred: or setup.cfg yea | 01:47 |
| mordred | dstufft: two things- a) the semantics of install_requires + dependency_links is confusing to people. the format in which that information is expressed in pip requirements files is clearer to devs that we've encountered | 01:47 |
| *** kaushikc has joined #openstack-dev | 01:47 | |
| dstufft | dependency_links, another thing on my list of things to murder :( | 01:48 |
| mordred | as in - a line with http://foo.bar.com/my.package.tar.gz#egg=my.pacakge for some reason makes sense to people | 01:48 |
| mordred | dstufft: please do - it's a batshit design | 01:48 |
| dstufft | mordred: so why is openstack using it? ;P | 01:48 |
| dstufft | My patch more or less kills dependency links I think, you'll need to use --allow-external PROJECTNAME to use them if they have a MD5 hash, or --allow-external PROJECTNAME --allow-insecure PROJECTNAME if they don't | 01:49 |
| mordred | dstufft: b) we have other tools that want to consume the depends list, and it turns out hte pip requirements file format is excellent for that (setup.cfg would also be excellent for it, stipulated) | 01:49 |
| mordred | dstufft: and c) we use a file of the same format to deal with test requirements, so that we can express what it takes to build a test env - and nothing in the current machienery supports this, really | 01:50 |
| mordred | dstufft: we need ot use specific links during dev cycle to be able to consume dev versions of libs we'll be releasing | 01:50 |
| dstufft | mordred: I normally just use a test extra for defining test requirements, (setuptools has test_requires btu it's horrible) | 01:50 |
| mordred | dstufft: also, d) we have 3 years of history that started with a rpo that did many worse things, and it's been a slow climb to something more sensible | 01:51 |
| dstufft | mordred: I work on PyPI, I'm well acquanted with d) :( | 01:52 |
| mordred | yah. d) is my favorite | 01:52 |
| mordred | but, I'll say, for the openstack dev workflow, having the requirements in an external file has been extremely nice | 01:53 |
| mordred | if I had my way about it, setuptools would learn to read an existing requirements.txt file if it existed by default - because it seems to be extremely discoverable by new devs | 01:53 |
| *** jbresnah has quit IRC | 01:53 | |
| mordred | setuptools is black magic to folks | 01:53 |
| mordred | requirements.txt is not | 01:54 |
| *** ogelbukh1 has quit IRC | 01:54 | |
| *** jbresnah has joined #openstack-dev | 01:54 | |
| *** kaushikc has quit IRC | 01:54 | |
| dstufft | mordred: are these new devs typically already Python devs? | 01:54 |
| *** maoy has quit IRC | 01:54 | |
| mordred | nope | 01:54 |
| mordred | we influx a rather large amount of new python devs all the time around here | 01:54 |
| mordred | the fun of being a project that is dragging hp and cisco and ibm from java into python :) | 01:55 |
| dstufft | So that's probably why then :) (and yes I agree requirements.txt is more discoverable than setup.py|cfg, but completely confusing to most experienced python devs :( ) | 01:55 |
| mordred | totaly | 01:55 |
| mordred | conceded | 01:55 |
| *** novas0x2a|laptop has joined #openstack-dev | 01:56 | |
| dstufft | mordred: I think there's a lot of things that could be done better here wrt packaging. I also think the current situation ends up being self refernetial as you need X thing because of Y thing, and you need Y thing because of Z thing, and you need Z thing because of X thing. I think the best path forward is probably to take the doc you'll hopefully send me, and I can work up a proposal, go back and forth and see if we can't hash out | 01:56 |
| dstufft | something that makes sense to both groups of people? | 01:56 |
| mordred | I think because of our influx like that, we tend to be more willing to question python community assumptions. sometimes we're right, sometimes we're wrong | 01:56 |
| mordred | dstufft: I'd love that | 01:56 |
| mordred | dstufft: and also, beer | 01:56 |
| mordred | dstufft: because what I want to wind up with is that default python packaging is sensible | 01:56 |
| mordred | and I have ZERO local hacks | 01:56 |
| dstufft | mordred: sure :) I don't drink but it doesn't bother me | 01:57 |
| * dstufft is an angry drunk | 01:57 | |
| mordred | dstufft: oh, well crap then. coffee/tea :) | 01:57 |
| * mordred is just a drunk | 01:57 | |
| * dstufft is also 6'4" and can do some hurts when angry | 01:57 | |
| mordred | eek. | 01:57 |
| * mordred makes note - no beer for dstufft | 01:57 | |
| dstufft | mordred: also +1 on default python packaging being sensible. It's finally headed that direction I think | 01:57 |
| mordred | cool | 01:58 |
| *** portante has joined #openstack-dev | 01:58 | |
| mordred | dstufft: also, I'll admit, we need to be better about communicating back out | 01:58 |
| mordred | we have some really fun operating semantics around here | 01:58 |
| dstufft | Even when some folks (death to distutils-sig) do everything they can to apply gratitous stop energy | 01:58 |
| mordred | and there are some things we need to support that other folks just usually don't | 01:58 |
| mordred | but without communication, our needs aren't going to get taken in to account, now are they? | 01:59 |
| mordred | dstufft: I'm also quite happy to help the cause, either with coding myself or with hiring people to code, if that's helpful | 01:59 |
| dstufft | mordred: PEP426 has gone out of it's way to support some pretty complex stuff. One of the main writers of it works for RedHat and has come at it from an angle of "how do I take this python package and turn itinto a RPM with the least amount of manual work" | 01:59 |
| dstufft | actually basically _the_ main writer of it at this point. The rest of it tend to just argue about what is or isn't a good idea | 02:00 |
| mordred | dstufft: for us, the main one is "how can I build a multi-project thing 100s of times a day" | 02:00 |
| mordred | s/100s/1000s/ | 02:00 |
| mordred | from scratch from a completely pristine new env | 02:01 |
| mordred | and how do I release without needing a sequence of versoin advancing commits :) | 02:01 |
| mordred | (the versioning code we use is also pretty important, and also completely not what setuptools expects/wants) | 02:01 |
| locke105 | setuptools ?= X.Y.Z | 02:02 |
| mordred | locke105: aroo? | 02:03 |
| locke105 | doesnt setuptools expect the X.Y.Z form | 02:03 |
| locke105 | which openstack uses | 02:03 |
| locke105 | 2013.1.1 | 02:03 |
| dstufft | ^ that isn't compat with PEP440 | 02:03 |
| uvirtbot | dstufft: Error: "that" is not a valid command. | 02:03 |
| dstufft | FWIW | 02:03 |
| mordred | dstufft: why not? | 02:03 |
| locke105 | oh right , 4 digit major version... | 02:04 |
| *** HenryG has quit IRC | 02:04 | |
| mordred | oh. there's a limit? | 02:04 |
| dstufft | mordred: date based releases are excluded, setuptools will consume it (and most tools will fall back to a less strict thing) | 02:04 |
| *** PaulM has quit IRC | 02:04 | |
| dstufft | IT's actually the first major digit being >= 1980 | 02:04 |
| mordred | dstufft: ignore the fact that 2013 is a year | 02:04 |
| dstufft | is the rule in PEP440 | 02:04 |
| mordred | the other numbers are not date based | 02:05 |
| mordred | well, that's going to be ablocker for us | 02:05 |
| mordred | because we can't verywell downgrade our existing versions now can we? | 02:05 |
| dstufft | http://www.python.org/dev/peps/pep-0440/ | 02:05 |
| mordred | sure. just saying | 02:05 |
| mordred | it's a rude rule | 02:05 |
| dstufft | mordred: well that's why they are excluded, because it makes it hard to go from that scheme to this scheme | 02:05 |
| *** anfrolov has joined #openstack-dev | 02:05 | |
| * clarkb does hate date based versions | 02:06 | |
| mordred | well, we work hard on making our versions mean something so less strict consumption should probably still work | 02:06 |
| clarkb | I have git for that | 02:06 |
| mordred | I don't care one way or the other about the first number being a year - it's just we're stuck with it | 02:06 |
| dstufft | (and semver semantics are heavily suggested with the version numbers, so it's suggested not to do breaking things in a 1.1 etc) | 02:06 |
| mordred | sure | 02:07 |
| mordred | we agree with that | 02:07 |
| mordred | well, kinda | 02:07 |
| * mordred was not consulted on the 2013.1 thing, btw | 02:07 | |
| mordred | anyway | 02:07 |
| mordred | locke105: the thing I'm talking about is how you set the fact that the version of the package is X.Y.Z | 02:07 |
| *** jhesketh_ has joined #openstack-dev | 02:07 | |
| dstufft | mordred: PEP440 isn't accepted yet so you should probably send nick a message | 02:08 |
| dstufft | either distutils-sig or uh, what's his email | 02:08 |
| mordred | I should. I mean, I suppor the intent there | 02:08 |
| dstufft | ncoghlan@gmail.com | 02:08 |
| mordred | I just hate being on the wrong side of a standardization with no way to get on the right side | 02:08 |
| dstufft | he's the Redhat guy whose doing the bulk of the writing | 02:08 |
| *** pmathews has joined #openstack-dev | 02:08 | |
| mordred | yeah, I know nick :) | 02:09 |
| locke105 | the clients changed version schemas i think | 02:09 |
| dstufft | mordred: There's some provisions for migration here, e.g. you can have a different package version point to a different upstream version | 02:09 |
| *** seanrob has joined #openstack-dev | 02:09 | |
| locke105 | they used to do the 2012.1 thing i think | 02:09 |
| mordred | they did. the distros shot me | 02:09 |
| locke105 | lol | 02:09 |
| *** seanrob has quit IRC | 02:09 | |
| mordred | locke105: not most of them | 02:09 |
| mordred | hrm | 02:09 |
| mordred | I'll talk to ttx | 02:09 |
| locke105 | yeah i guess that got changed before many of them came along | 02:09 |
| *** seanrob has joined #openstack-dev | 02:09 | |
| clarkb | mordred versions should be meaningful. dates are just redundant | 02:09 |
| dstufft | so you could have 13.1.1 on PyPI land point to 2013.1.1 using the source_label | 02:09 |
| mordred | it's a good point you bring up | 02:09 |
| mordred | dstufft: evil evil | 02:10 |
| dstufft | mordred: yea :/ | 02:10 |
| mordred | dstufft: that would just sow more confusion | 02:10 |
| dstufft | It's not a good place, and the problem with just allowing 2013 is that it makes it impossible to go from 2013 -> 2.0 (or whatever) | 02:10 |
| dstufft | which is why PEP440 specifically excludes it | 02:10 |
| mordred | yeah | 02:10 |
| mordred | no, I follow | 02:11 |
| mordred | also, I actually care less about that | 02:11 |
| dstufft | :) | 02:11 |
| *** jecarey has quit IRC | 02:11 | |
| mordred | but you make a good point about the fact that we're not communicating sensibly about major version breakage | 02:11 |
| dstufft | I forget sometimes who is yelling at me about what new thing :D | 02:11 |
| locke105 | wasn't PBR supposed to fix the version setup stuff | 02:11 |
| mordred | locke105: it does | 02:11 |
| locke105 | sorry pbr | 02:11 |
| dstufft | pytz got upset about that rule :( | 02:11 |
| mordred | locke105: it works GREAT | 02:11 |
| dstufft | mordred: I actually think pip 1.4 might not install 2013.1.1 by default | 02:11 |
| mordred | locke105: I'm saying, that's one of the things we do that's kinda openstack-specific | 02:12 |
| mordred | dstufft: in favor of what? | 02:12 |
| dstufft | if you do ``pip install foo`` | 02:12 |
| mordred | oh god | 02:12 |
| mordred | really? | 02:12 |
| dstufft | I forget if distlib can parse that or not | 02:12 |
| dstufft | lemme look | 02:12 |
| *** seanrob has quit IRC | 02:12 | |
| * locke105 is curious now too | 02:12 | |
| mordred | that might be a bit of a major nightmare for us | 02:12 |
| mordred | oh. nope | 02:12 |
| mordred | we'll be fine | 02:12 |
| *** seanrob has joined #openstack-dev | 02:12 | |
| mordred | I thought markmc had persisted in keeping date versions for oslo.config | 02:13 |
| dstufft | ah no I'm wrong | 02:13 |
| dstufft | distlib can parse 2013.1.1 | 02:13 |
| mordred | dstufft: the only thing we use date versioning on is stuff we don't publish to pypi anyway | 02:13 |
| mordred | like nova | 02:13 |
| dstufft | mordred: pip 1.4 excludes "pre release" things by default | 02:14 |
| locke105 | if only pypi had 'epoch' numbers like rpm | 02:14 |
| mordred | dstufft: that makes me super happy | 02:14 |
| mordred | dstufft: and will simplify a few things we do | 02:14 |
| dstufft | requiring you to add either --pre to the install comamnd, or to include a prerelease int he version specifier (==1.0.dev1, >=0.0.0.dev0) etc, doesn't include != ) | 02:14 |
| dstufft | https://github.com/pypa/pip/blob/develop/pip/util.py#L678-L692 is the function that determiens pre release or not | 02:14 |
| dstufft | mordred: it breaks pytz which is why I thought it might break 2013.1.1 (pytz uses 2013b, 2013f, etc) | 02:16 |
| mordred | dstufft: yeah, so if we dod ==1.0.0a1, we can upload that to pypi post pip 1.4 and pip install foo will still just get the one before | 02:16 |
| dstufft | mordred: yup | 02:16 |
| mordred | dstufft: I'm even more in favor of 1.4 being released | 02:16 |
| mordred | can haz now? | 02:16 |
| *** nati_ueno has joined #openstack-dev | 02:16 | |
| dstufft | mordred: https://github.com/pypa/pip/issues?milestone=1&state=open | 02:16 |
| mordred | dstufft: because right now we use tarballs.openstack.org as a place to put pre-release stuff, but that's super kludgy | 02:16 |
| clarkb | ooh I want pip 1.4 too | 02:17 |
| mordred | and just having a pure build rule of "it got a tag, shove it on pypi" would be neat | 02:17 |
| clarkb | that pre release stuff will be very helpful | 02:17 |
| dstufft | mordred: obviously folks who haven't upgraded to to 1.4 will install it by default | 02:17 |
| dstufft | and who the fuck knows what easy_install and buildout is doing | 02:17 |
| dstufft | but I can only deal with so many projects at once :D | 02:17 |
| locke105 | lol | 02:17 |
| mordred | dstufft: I don't care about any of those people | 02:17 |
| *** nati_ueno has quit IRC | 02:18 | |
| mordred | dstufft: I'm stuck with many things, caring about every permutation of what we might could support is thankfully not one of them | 02:18 |
| dstufft | mordred: some of that milestone stuff might get dropped, but ATM that's stuff that we (we being the pip maintainers) are trying to target to get into 1.4 | 02:18 |
| mordred | ossum | 02:19 |
| dstufft | mordred: what are you goys using to mirror PyPI btw? | 02:19 |
| mordred | dstufft: we have some scripts... one sec | 02:19 |
| dstufft | Becasue there's a good chance my CDN broke it if you're using an older version of the normal mirroring stuff | 02:19 |
| locke105 | pypimirror i think | 02:19 |
| dstufft | (many tears were had, some fixes made) | 02:20 |
| clarkb | dstufft: no no the upstream stuff is bad | 02:20 |
| *** PaulM has joined #openstack-dev | 02:20 | |
| clarkb | it doesn't actually create a mirror because of the external links | 02:20 |
| mordred | nonono | 02:20 |
| mordred | we don't mirror pypi | 02:20 |
| mordred | we build our own mirror of just the stuff we use | 02:20 |
| clarkb | we actually install everything then capture our PIP_CACHE and use that in the mirror | 02:20 |
| mordred | https://github.com/openstack-infra/jeepyb/blob/master/jeepyb/cmd/run_mirror.py | 02:20 |
| dstufft | oh | 02:20 |
| dstufft | gotcha | 02:20 |
| clarkb | otherwise you may just mirror a link and that is useless | 02:20 |
| *** armax has joined #openstack-dev | 02:20 | |
| dstufft | yea there's a thing by hoelger that mirrors the external links too | 02:21 |
| dstufft | (can't spell name) | 02:21 |
| *** sride has quit IRC | 02:21 | |
| clarkb | and we gate updates to the mirror by checking that everything can actually install together | 02:21 |
| Alex_Gaynor | holger* | 02:21 |
| clarkb | so that is how we tackle the example you had before about requirements that don't jive with each other | 02:22 |
| dstufft | (I just wanted to make sure you guys didn't end up with a borken mirror due to CDN if you were using the changelog api) | 02:22 |
| clarkb | we don't allow them at all | 02:22 |
| locke105 | mordred: so this just basically serves a pip cache over HTTP? | 02:22 |
| dstufft | (also changelog api is bad) | 02:22 |
| clarkb | locke105: basically | 02:22 |
| locke105 | interesting | 02:23 |
| locke105 | you can use just HTTPd like apache or nginx then? | 02:23 |
| Alex_Gaynor | Not sure who to talk to about this, but if http://pypi-externals.caremad.io/help/what/ could be done for the openstack packages like the clietns and tools and whatnot that would be awesome | 02:23 |
| clarkb | locke105: yes we put apache with mod_index in front of it | 02:23 |
| dstufft | Alex_Gaynor does a better job of sending my link around then I do +1 | 02:23 |
| locke105 | clarkb: cool, didn't realize it was that simple to make a pypi index | 02:23 |
| dstufft | locke105: a PyPI index is more complex, but pip and such can function with a simpler one | 02:24 |
| locke105 | oh | 02:24 |
| dstufft | (only slightly more complex though) | 02:24 |
| locke105 | go pip | 02:24 |
| clarkb | Alex_Gaynor: dstufft I am being dense but what does that link describe? | 02:24 |
| dstufft | clarkb: you know those external urls you hate on PyPI? | 02:25 |
| clarkb | dstufft: yes | 02:25 |
| Alex_Gaynor | clarkb: basically if you follow those instructions for packages you have on pypi, stuff installs faster for yoru users | 02:25 |
| mordred | oh, yeah. I keep meaning to do that | 02:25 |
| dstufft | clarkb: I deployed a change 2 weeks ago or so that lets package authors remove all of them from their package | 02:25 |
| dstufft | that link tells you how to do it | 02:25 |
| mordred | is there a pypi rest call for htat? | 02:25 |
| clarkb | dstufft: and pypi actually mirrors them sanely if you do that? | 02:25 |
| clarkb | mordred: REST? hahaha | 02:25 |
| clarkb | mordred: they do have na http of ssh interface :P | 02:25 |
| mordred | cause if so, we could get it done for all of our projects really easy | 02:25 |
| *** ayoung has quit IRC | 02:26 | |
| dstufft | mordred: There's a form you can submit and it doesn't have CSRF :V | 02:26 |
| dstufft | or bug me and I can issue a query | 02:26 |
| mordred | or apiu call | 02:26 |
| mordred | don't care if it's rest :) | 02:26 |
| mordred | dstufft: well, we have project creation and management scripted | 02:26 |
| mordred | from a yaml file | 02:26 |
| mordred | so, what I'd love to do is add in support for the form submission | 02:26 |
| clarkb | dstufft: does that break things that depend on external links or do you actually mirror those packages? | 02:27 |
| dstufft | clarkb: it still requires uploading to PyPI directly, but it removes the external urls so that even if you have files uploaded, pip doesn't need to scrape external files | 02:27 |
| mordred | so we can make sure to continue to set it ongoing | 02:27 |
| mordred | dstufft: we upload to pypi directly already quite nicely :) | 02:27 |
| mordred | so, yeah, we want that big time | 02:27 |
| dstufft | clarkb: example http://pypi-externals.caremad.io/python-keystoneclient/ pip has to fetch 3 urls just to find what versions are available | 02:27 |
| mordred | we're already good upload citizens | 02:27 |
| *** alop has quit IRC | 02:27 | |
| clarkb | dstufft: oh yeah because it was broken before. Ok less dense now | 02:28 |
| dstufft | Begining of july I'll be auto migrating ~25k projects to the new method | 02:28 |
| clarkb | dstufft: can I ask what was the motivation behind http over ssh? | 02:28 |
| mordred | dstufft: are you going to suck in their existing external files? | 02:29 |
| dstufft | clarkb: because of bad people who wouldn't shut up that CACert is a "real" certificate | 02:29 |
| dstufft | (the same bad people implmeneted http over ssh) | 02:29 |
| dstufft | mordred: No, we can't legally do that. But the 25k i'm migrating are tested to not require it already | 02:29 |
| mordred | ossum | 02:29 |
| dstufft | I wrote a script that scraped all of PyPI to determine what versions are available for every package | 02:29 |
| dstufft | and from what urls | 02:30 |
| mordred | well, anyting that has the owner openstackci is good to go | 02:30 |
| *** rongze has joined #openstack-dev | 02:30 | |
| dstufft | so we could determine who it was safe to turn it off for | 02:30 |
| dstufft | mordred: cool, if you promise not to yell at me I can do that :) | 02:31 |
| mordred | dstufft: go for it! | 02:31 |
| mordred | dstufft: we do not believe in external links | 02:31 |
| dstufft | https://github.com/dstufft/pypi.linkcheck/tree/migration is the scripts and data files fwiw | 02:31 |
| dstufft | incase you were interested! | 02:31 |
| mordred | clarkb: I do think that we should add a form submission to manage_projects | 02:32 |
| mordred | although... | 02:32 |
| mordred | dstufft: post july is it going to be the default? | 02:32 |
| dstufft | It's already the default for newly registered projects | 02:32 |
| clarkb | mordred: to setup pypi? | 02:32 |
| mordred | dstufft: because if so, implemeting automation around ensuring that it's set on our projects .. | 02:32 |
| mordred | cool | 02:32 |
| mordred | clarkb: nevermind. we don't need to do it | 02:32 |
| clarkb | mordred: pypi doesn't have a real API which makes that annoying | 02:32 |
| dstufft | and has been for a few weeks since I deployed the code | 02:32 |
| mordred | cool | 02:32 |
| clarkb | mordred: instead its go read cheeseshop code, curse http over ssh, use curl insecurely | 02:33 |
| mordred | I think that a bulk switch of anything that's owned or managed by openstackci | 02:33 |
| mordred | should get us | 02:33 |
| *** sld_ is now known as sld | 02:35 | |
| *** seanrob has quit IRC | 02:35 | |
| *** seanrob has joined #openstack-dev | 02:35 | |
| locke105 | why 'jeepyb' | 02:36 |
| locke105 | is there a reference i'm not aware of? | 02:36 |
| clarkb | locke105: Gerrit Project Builder -> GPB -> jeepyb | 02:36 |
| locke105 | ah... | 02:37 |
| locke105 | it all makes sense now :) | 02:37 |
| Alex_Gaynor | dstufft, mordred, clarkb: FWIW I just confirmed that no projects owned by openstackci rely on externally hosted stuff. | 02:37 |
| clarkb | \o/ | 02:37 |
| clarkb | Alex_Gaynor: thank you for checking | 02:37 |
| Alex_Gaynor | clarkb: dstufft wrote a script for it, all I did was run it :) | 02:37 |
| dstufft | mordred: https://gist.github.com/dstufft/5754154 verify this list looks ok? | 02:39 |
| openstackgerrit | A change was merged to openstack/nova: DB migration to the new BDM data format https://review.openstack.org/29251 | 02:39 |
| *** cp16net is now known as cp16net|away | 02:40 | |
| *** cp16net|away is now known as cp16net | 02:40 | |
| openstackgerrit | A change was merged to openstack/nova: Create an image BDM for every instance https://review.openstack.org/29252 | 02:40 |
| mordred | yup | 02:40 |
| mordred | dstufft: looks great! | 02:40 |
| clarkb | dstufft: Alex_Gaynor: I don't think it is mentioned yet, but if you guys wanted to watch openstack-dev/pbr on review.openstack.org I think that would be awesome | 02:41 |
| *** otherwiseguy has quit IRC | 02:48 | |
| *** noslzzp has quit IRC | 02:48 | |
| *** markvoelker has quit IRC | 02:48 | |
| dstufft | clarkb: how I do | 02:48 |
| dstufft | clarkb: My stuff @ work is shielded from upstream for the most part so i've never done :) | 02:49 |
| clarkb | dstufft: https://review.openstack.org/#/settings/projects if you go there you can add watches | 02:49 |
| clarkb | add openstack-dev/pbr to the project name text box, click watch, then select how you want to be notified of changes in the list below | 02:49 |
| dstufft | ok done | 02:50 |
| clarkb | great thanks | 02:50 |
| *** otherwiseguy has joined #openstack-dev | 02:50 | |
| dstufft | mordred: Alex_Gaynor openstack updated | 02:50 |
| dstufft | http://pypi-externals.caremad.io/python-keystoneclient/ eample | 02:50 |
| clarkb | (this has been a good conversation and don't want it to be a one time thing) | 02:51 |
| Alex_Gaynor | It now takes less than half as long to scrape all those pages, yay | 02:52 |
| clarkb | dstufft: mordred just pushed updates to some pbr changes so you should've gotten ntofications maybe | 02:52 |
| mordred | Alex_Gaynor: yay! | 02:52 |
| dstufft | yea I got them | 02:52 |
| mordred | woot | 02:52 |
| dstufft | now I need to figure out how this gerrit thing works | 02:52 |
| Alex_Gaynor | dstufft: you reviewed a patch of mine once, figure it out :) | 02:52 |
| mordred | dstufft: tell anotherjesse to let you work upstream more... | 02:53 |
| clarkb | dstufft: people tend to resist it at first but after using it for a couple changes its like the best thing ever | 02:53 |
| dstufft | Alex_Gaynor: press buttans until thing happens | 02:53 |
| dstufft | that's how I computer | 02:53 |
| locke105 | beep boop | 02:53 |
| clarkb | didn't you hear github is the small guy in this battle :) | 02:53 |
| Alex_Gaynor | I thought jesse was no longer at Nebula? | 02:53 |
| dstufft | so why does everyone say someone else must approve all the time, that confused me when I reviewed Alex_Gaynor's thing | 02:54 |
| Alex_Gaynor | dstufft: its based on number of core reviewers, not total reviewers | 02:54 |
| Alex_Gaynor | (IIUC) | 02:54 |
| dstufft | oh | 02:54 |
| clarkb | yeah the numeric values gerrit uses don't really represent things as well | 02:54 |
| clarkb | its not regular math. Instead think of a +1 code review as a thumbs up and a +2 code review as a gold star | 02:55 |
| *** markvoelker has joined #openstack-dev | 02:55 | |
| mordred | yeah. google got a little bit to clever on those I think | 02:55 |
| clarkb | you need two gold stars before a core member can approve/merge the code regardless of the number of thumbs up | 02:55 |
| lifeless | its really set based | 02:55 |
| *** kaushikc has joined #openstack-dev | 02:55 | |
| lifeless | 4 *+1 = +1 | 02:55 |
| mordred | Alex_Gaynor: really? I didn't realize he's left | 02:55 |
| mordred | he'd | 02:55 |
| Alex_Gaynor | mordred: I assumed so, he was contracting at the company I was at before I joined rackspace (rdio) | 02:56 |
| clarkb | and a -2 is a giant red F on that term paper | 02:56 |
| mordred | looks like you're right | 02:56 |
| clarkb | mordred: the amount of pbr fail in the zuul queue right now is too damn high :) | 02:57 |
| mordred | clarkb: yeah, I know.I can't reproduce that issue locally | 02:57 |
| mordred | clarkb: so I'm poking jenkins in the face | 02:57 |
| *** souvik has joined #openstack-dev | 02:57 | |
| *** pmathews has quit IRC | 02:58 | |
| dstufft | mordred: fwiw I have a ton of irons in the fire wrt packaging so if I forget stuff you want/need here just yell at me | 02:58 |
| *** mjfork has quit IRC | 02:58 | |
| mordred | dstufft: totes | 02:58 |
| mordred | dstufft: and thanks | 02:59 |
| *** zaneb has quit IRC | 02:59 | |
| locke105 | does anyone know if review.openstack.org is a single host Gerrit install or is run with read-only slave nodes, etc? | 02:59 |
| mordred | locke105: single host | 02:59 |
| locke105 | i have one set up at work but apparently its hard to make it highly available | 03:00 |
| mordred | locke105: plus a git mirror served by apache | 03:00 |
| locke105 | ? | 03:00 |
| mordred | locke105: oh yeah. HA gerrit != fun | 03:00 |
| locke105 | git mirror | 03:00 |
| mordred | locke105: we replicate git repos to a local location that we serve via apache, so requests to some of the gerrit urls don't have to hit java | 03:00 |
| locke105 | so you have it replicate to another server and then use HTTP to serve the git requests | 03:00 |
| mordred | but that's a scaling thing | 03:01 |
| locke105 | oh k | 03:01 |
| mordred | actually, same server :) | 03:01 |
| locke105 | lol | 03:01 |
| mordred | it's to offload java thread issues | 03:01 |
| *** kaushikc has quit IRC | 03:01 | |
| *** kaushikc has joined #openstack-dev | 03:01 | |
| clarkb | apparently android does specail gerrit sauce for ha that isn't public | 03:01 |
| locke105 | i read in the change log for 2.6(or 2.7?) that the internal JGit daemon is "superfast and faster than CGit's daemon" | 03:01 |
| clarkb | or maybe its just google | 03:01 |
| clarkb | but yeah HA gerrit makes my brain hurt | 03:01 |
| mordred | locke105: it might be - but it consumes the same internal thread and memory pool as the rest of gerrit | 03:02 |
| mordred | HA gerrit should be fairly simple with DRBD | 03:02 |
| locke105 | yeah... | 03:02 |
| mordred | anything more complex like an active/active setup would be a bit more of a nightmare | 03:02 |
| locke105 | yes they say its not too hard to do a master - slave failover | 03:02 |
| *** ayoung has joined #openstack-dev | 03:02 | |
| clarkb | mordred: an object store would probably do it reasonably well | 03:02 |
| clarkb | (active active) | 03:02 |
| mordred | clarkb: you still need to deal with mysql | 03:02 |
| locke105 | tbh | 03:03 |
| mordred | but if you offload that and consider that a separate engineering task | 03:03 |
| locke105 | thats not really the hard part, | 03:03 |
| mordred | what isn't? | 03:03 |
| locke105 | its the stupid amount of caching in Gerrit | 03:03 |
| locke105 | mysql | 03:03 |
| clarkb | caching is fine | 03:03 |
| *** PaulM has quit IRC | 03:03 | |
| * mordred used to be an HA consultant for MySQL - totally agrees | 03:03 | |
| *** jamielennox is now known as jamielennox|away | 03:03 | |
| clarkb | its resolving git atomic operations | 03:03 |
| mordred | caching isn't fine if you want active/active | 03:03 |
| locke105 | yes gerrit has super long cache timeouts | 03:04 |
| locke105 | its not really a cache IMO | 03:04 |
| mordred | but I'd be happy with active/passive failover | 03:04 |
| locke105 | true | 03:04 |
| locke105 | definitely a step up | 03:04 |
| mordred | or - if I get a little bit more headcount - I want to get someone to actually put HA support into gerrit | 03:04 |
| mordred | (that doesn't depend on magic google sauce) | 03:04 |
| locke105 | its a hugely requested feature now | 03:04 |
| mordred | yah | 03:05 |
| *** seanrob has quit IRC | 03:05 | |
| clarkb | where is gerrit caching that is the problem? | 03:05 |
| locke105 | i'm not sure if my java would be up to par... | 03:05 |
| * locke105 goes to find the thread for Gerrit HA | 03:05 | |
| clarkb | I don't really care if gerrit show reviewers old patchsets or if git fetches pull old stuff. its the gerrit merge this thing please without making your buddy unhappy that bothers me | 03:06 |
| mikal | mordred: last week I was locked in a hospital caring for a family member (who is now ok). I got bored and did this -- http://openstack.stillhq.com/ci/I932a64df5847bc26ef321a35fadc0b60fc4b317f/1/sqlalchemy_migration_nova_user_001/log.html | 03:07 |
| *** souvik has quit IRC | 03:07 | |
| mikal | mordred: do you think that's useful? | 03:07 |
| *** kaushikc1 has joined #openstack-dev | 03:07 | |
| mordred | mikal: yah | 03:08 |
| mordred | mikal: also, glad your family member is now ok | 03:08 |
| mikal | mordred: its kind of basic, but there's only so much you can do on 3g | 03:08 |
| mikal | mordred: thanks | 03:08 |
| locke105 | clarkb: gerrit HA thread https://groups.google.com/group/repo-discuss/tree/browse_frm/month/2011-02/fe8501de25c68bee?rnum=51&_done=/group/repo-discuss/browse_frm/month/2011-02?&pli=1 | 03:08 |
| locke105 | kind of old now | 03:08 |
| *** kaushikc has quit IRC | 03:09 | |
| *** jbresnah has quit IRC | 03:09 | |
| clarkb | ah yeah that would be annoying | 03:09 |
| locke105 | can get around some of it by setting the cache expiration times really low | 03:10 |
| locke105 | i am pretty sure that one needs a shared filesystem for the Git repos still | 03:11 |
| *** zaneb has joined #openstack-dev | 03:11 | |
| *** Mandell has joined #openstack-dev | 03:11 | |
| clarkb | swift! someone did write an AWS backend for git | 03:12 |
| clarkb | I wonder how slow that is | 03:13 |
| *** prad has quit IRC | 03:13 | |
| clarkb | and if aggressively packing helps | 03:13 |
| mordred | ceph ftw | 03:13 |
| locke105 | sounds interesting anyways | 03:13 |
| mordred | ceph-backed-afs! | 03:13 |
| mordred | s/ceph/rados/ | 03:13 |
| mordred | whatev | 03:13 |
| * mordred shuts up | 03:14 | |
| dstufft | mordred: heh, ceph is on my list of things to look at for PyPI :/ | 03:14 |
| dstufft | HA'ing FS access ftl | 03:14 |
| *** armax has quit IRC | 03:14 | |
| mordred | dstufft: do you know florian? | 03:14 |
| dstufft | mordred: I know someone named florian, I doubt it's the same person :) | 03:15 |
| clarkb | seems like an object store (nevermind FS) would work well for pypi | 03:15 |
| mikal | mordred: you should fix https://review.openstack.org/#/c/29611/ | 03:15 |
| dstufft | clarkb: that's the goal, right now i depends heavily on the FS | 03:15 |
| dstufft | it* | 03:15 |
| mordred | dstufft: florian haas. works for hastexo doing ceph and pacemaker stuff | 03:15 |
| dstufft | and no unit tests :V | 03:15 |
| mordred | also does a lot with openstack | 03:15 |
| dstufft | mordred: nope don't know him | 03:15 |
| locke105 | pypi is kind of just a case specific object store.... | 03:15 |
| mordred | well - if you want to go object-store, I can getcha a free cloud-based one in a hurry I'm pretty sure ;) | 03:16 |
| dstufft | mordred: ops czar isn't a fan of outsourcing stuff we can reasonably host in house | 03:17 |
| mordred | oh well. I'm a fan of outsourcing everything I possibly can | 03:17 |
| dstufft | (I don't particularly care either way, but I'm not the decision maker) | 03:18 |
| mordred | dstufft: openstack infra runs 100% on public clouds | 03:18 |
| mordred | we have zero machines | 03:18 |
| dstufft | PSF has a bunch of machines running inside of OSUOL with VM's running ontop of them | 03:18 |
| mordred | but - yeah - I'm not your ops czar - so, really, whatever he wants to do I guess | 03:18 |
| clarkb | locke105: how so? naively its just a bunch of packages (objects) that need to be referred to and retrieved | 03:18 |
| * locke105 wishes he could use public clouds for work. | 03:19 | |
| locke105 | yes | 03:19 |
| clarkb | yay OSUOSL | 03:19 |
| locke105 | clarkb: i was implying that they are very similar | 03:19 |
| locke105 | put once, read many type storage | 03:19 |
| dstufft | yea | 03:20 |
| clarkb | ah I read it as needing special ness for some reason | 03:20 |
| locke105 | pypi has some nice presentation of the metadata | 03:20 |
| dstufft | The goal is to get it on some type of object store eventually | 03:20 |
| mordred | clarkb: when are you going to get OSUOSL to install an openstack | 03:20 |
| mordred | ? | 03:20 |
| clarkb | mordred: they run ganeti | 03:20 |
| clarkb | and lance seems reasonably happy with it | 03:20 |
| dstufft | ah yea ganeti is the thing we have there I think | 03:20 |
| clarkb | mordred: PSU is apparently going to try a test openstack install which makes me happy | 03:20 |
| dstufft | maybe we don't have it I dunno, maybe it's their ifnra | 03:20 |
| mordred | clarkb: cool | 03:20 |
| * dstufft just tries not to blow shit up | 03:20 | |
| clarkb | dstufft: but blowing stuff up makes the lives of undergraduate sysadmins interesting | 03:21 |
| locke105 | sound strategy | 03:21 |
| dstufft | mordred: there, I just merged https://github.com/pypa/pip/pull/985 so it'll for sure be in 1.4 | 03:21 |
| *** boris-42 has joined #openstack-dev | 03:21 | |
| mordred | woot | 03:21 |
| clarkb | mordred: we could do an OSUOSL counterpart with our free clouds and make our providers fall over | 03:22 |
| Alex_Gaynor | dstufft: will use of insecure externals warn? | 03:22 |
| dstufft | Alex_Gaynor: ya | 03:22 |
| Alex_Gaynor | yay | 03:22 |
| Alex_Gaynor | dstufft: do they have a timetable for the 1.4 relase? | 03:22 |
| dstufft | Alex_Gaynor: more or less when this stuff is done https://github.com/pypa/pip/issues?milestone=1&state=open | 03:22 |
| dstufft | assuming we don't table anything | 03:22 |
| dstufft | Alex_Gaynor: warning https://gist.github.com/dstufft/f2d5b8fef8a0e66f3d77#file-immediate_behavior-L41-L42 | 03:23 |
| Alex_Gaynor | cool | 03:24 |
| Alex_Gaynor | dstufft: you should make the warnings have colors | 03:24 |
| dstufft | that'll be the first use of color in pip, I'll need to sort out how to make it not bad in various situations the output could end up in | 03:24 |
| dstufft | not a bad idea though | 03:24 |
| *** russellb has quit IRC | 03:26 | |
| locke105 | i feel like Gerrit has become a lot more popular due to OpenStack using it | 03:28 |
| locke105 | guess that could be said of a lot of things that OpenStack uses... | 03:29 |
| mordred | we're fun like that | 03:30 |
| mordred | dstufft: we've got code somewhere that used color and then avoids using it when in a pipeline... | 03:31 |
| locke105 | TBH though, i wonder how openstack velocity would be affected by not having Gerrit | 03:31 |
| mordred | locke105: quite badly I believe | 03:31 |
| mordred | locke105: especially once you consider zuul as well | 03:31 |
| locke105 | mordred: yeah thats true | 03:31 |
| *** russellb has joined #openstack-dev | 03:32 | |
| *** vishal has joined #openstack-dev | 03:33 | |
| *** alunduil has quit IRC | 03:34 | |
| *** alunduil has joined #openstack-dev | 03:35 | |
| clarkb | zuul helps tremendously | 03:37 |
| locke105 | clarkb: yes, but i think the interesting part is in the part where Zuul is basically applying superscalar technique to jenkins jobs and Gerrit change sets | 03:38 |
| locke105 | and its even necessary now to handle the amount of merges | 03:38 |
| clarkb | over the last 24 hours we put ~937 tests through the gate queue | 03:39 |
| clarkb | unfortunately logstash doesn't make it easy to get a count of changes at the moment | 03:40 |
| * locke105 thinks that is f'n crazy. | 03:40 | |
| *** marun has joined #openstack-dev | 03:40 | |
| clarkb | locke105: ~3671 total tests. Thats actually a semi average day. I think we have done well over 5k during release/milestone weeks | 03:41 |
| clarkb | ~160 million log lines that we are indexing per day. Sorry logstash/kibana are the new shiny toy so I enjoy grabbing numbers out of it | 03:41 |
| locke105 | so... | 03:42 |
| locke105 | does anyone know if a CI system that has to deal with more merges/changes than that.... | 03:42 |
| locke105 | daily | 03:42 |
| dstufft | Travis-ci doesn't give any numbers but they might | 03:43 |
| *** kaushikc1 has quit IRC | 03:43 | |
| clarkb | travis does a non trivial amount of testing. based on their 2012 december numbers I roughly approximated we would be 1/12 of their load if we used them | 03:43 |
| clarkb | dstufft: they did once. but also weren't very specific | 03:43 |
| dstufft | oh yea I forgot about that | 03:44 |
| locke105 | wow | 03:44 |
| clarkb | http://about.travis-ci.org/blog/2012-12-17-numbers/ | 03:45 |
| *** lonelyibex has quit IRC | 03:46 | |
| lifeless | test-minutes would be a more interesting number | 03:46 |
| locke105 | ya | 03:46 |
| *** SergeyLukjanov has joined #openstack-dev | 03:47 | |
| clarkb | definitely | 03:47 |
| mordred | also, I do not think that any of travis's tests are like the devstack gate :) | 03:48 |
| clarkb | probably not | 03:49 |
| clarkb | also http://logstash.openstack.org exists as does graphite.openstack.org if people want to mine out interesting info | 03:49 |
| mordred | like github, I think travis serves its intended market extremely well, and I'm super happy that they are providing ci for normal folks | 03:49 |
| clarkb | I would be super excited if people randomly showed up on IRC and were all "look at the cool info we got out of logstash/graphite" | 03:50 |
| *** dims has quit IRC | 03:50 | |
| mordred | clarkb: ++ | 03:50 |
| locke105 | TBH, i was not aware of travis-ci before this convo, and i find myself wondering why we have internal build teams.... | 03:50 |
| clarkb | making logstash do 160 million events per day has been a fair amount of work and I want to see it used for somethign :) | 03:50 |
| locke105 | s/we/at work/ | 03:50 |
| mordred | locke105: well, there should never be a thing called a "build team" | 03:50 |
| *** neelashah has joined #openstack-dev | 03:51 | |
| locke105 | yeah.... | 03:51 |
| lifeless | locke105: travis-ci has very unpredictable latency IME | 03:51 |
| mordred | because there is literally nothing they can do better than automation | 03:51 |
| mordred | now, a build automation team or a CI team can be super helpful | 03:51 |
| clarkb | lifeless: yeah, the free stuff is round robinned across a static set of resources | 03:51 |
| mordred | but depending on scale, outsourcing those things is a fine choice too | 03:51 |
| *** koolhead17 has joined #openstack-dev | 03:51 | |
| clarkb | lifeless: I imagine their non free stuff is much better | 03:52 |
| mordred | one would imagine | 03:52 |
| mordred | :) | 03:52 |
| dstufft | clarkb: they've done some improvements in that area with QoS stuff so one project doesn't chew up all the avail builders | 03:52 |
| locke105 | interesting | 03:52 |
| dstufft | I have the paid thing, it's better/more consistent | 03:52 |
| *** ogelbukh1 has joined #openstack-dev | 03:52 | |
| dstufft | but can be pricey, think I paid $69/month for 1 concurrent build | 03:53 |
| lifeless | hah | 03:53 |
| lifeless | you can get a full VM in HPCS for that | 03:53 |
| lifeless | more tan | 03:53 |
| lifeless | 3 I think | 03:53 |
| dstufft | yea but then I need to manage it :) | 03:53 |
| vishy | smoser: nothing useful in there | 03:53 |
| vishy | smoser: i wonder if this would be helpful https://dev.openwrt.org/ticket/5299 | 03:54 |
| lifeless | dstufft: shrug | 03:54 |
| lifeless | dstufft: since we're a massive automation project | 03:54 |
| lifeless | dstufft: I kinda figure everyone here has automation mindset :) | 03:54 |
| * mordred has free clouds ... | 03:55 | |
| lifeless | vishy: that looks siny | 03:55 |
| lifeless | vishy: I was wishing we had that facility in stock cloud images the other day | 03:55 |
| * mordred hasn't thought about the cost of the resources he consumes like a rabit beast in a while | 03:55 | |
| lifeless | shiny[6~ | 03:55 |
| dstufft | lifeless: someone needs to monitor/manage the automation ;P | 03:55 |
| lifeless | mordred: s/d/t/ | 03:55 |
| * locke105 has a habit of almost automating himself out of jobs. | 03:55 | |
| vishy | lifeless: it works fine in most images | 03:56 |
| vishy | lifeless: just not in cirros | 03:56 |
| mordred | locke105: we have found that automatoin here breeds more automation. not less :) | 03:56 |
| lifeless | vishy: oh! didn't realise it was in the default ruleset already | 03:56 |
| lifeless | I can make baremetal more shiny! | 03:56 |
| mordred | dstufft: indeed. we have several full time staff (and need more) managing the openstack build automation | 03:56 |
| vishy | smoser: http://paste.openstack.org/show/38341/ | 03:56 |
| vishy | lifeless: yeah you can push out default routes with dhcp | 03:56 |
| vishy | useful with quantum | 03:56 |
| locke105 | mordred: yeah but openstack has no profit margin | 03:56 |
| lifeless | vishy: wait; we may be talking different things. | 03:57 |
| lifeless | vishy: I'm talking non-default routes. | 03:57 |
| vishy | lifeless: nope we aren't | 03:57 |
| lifeless | vishy: default is supported approx everywhere. | 03:57 |
| lifeless | vishy: non-default is different | 03:57 |
| vishy | lifeless: yeah you can push out a gateway no problem | 03:57 |
| *** sdake_ has joined #openstack-dev | 03:57 | |
| vishy | this is separate | 03:57 |
| lifeless | ok good | 03:57 |
| vishy | pushing out a static route via dhcp | 03:57 |
| mordred | locke105: openstack itself has no profit at all - however, most of the peope managing openstack automation do work for for-profit companies :) | 03:57 |
| mordred | locke105: it's a fun combo | 03:57 |
| *** comay has quit IRC | 03:57 | |
| clarkb | 65 gate-tempest-devstack-vm-full test runs that completed in the zuul gate queue | 03:57 |
| mordred | clarkb: wow. slow day | 03:58 |
| clarkb | each of those tests is ~45 minutes and we run two to three other tempest tests :) | 03:58 |
| mordred | locke105: otoh - there are actually some great papers on the ROI around CI/automation investment | 03:58 |
| vishy | quantum does it for 169.254/16 if you use the enable_isolated_metadata | 03:59 |
| mordred | locke105: there are good numbers you can put on it, even on a fully for-profit/have-margins project to justify spending wisely in the area | 03:59 |
| vishy | lifeless: ^^ | 03:59 |
| clarkb | mordred: like being able to potentially merge 65 changes in a day serialishly without running out of hours in the day? | 03:59 |
| mordred | clarkb: there's that :) | 03:59 |
| locke105 | mordred: i don't doubt CI, i think its great | 03:59 |
| lifeless | vishy: cool | 03:59 |
| mordred | http://www.rbcs-us.com/images/documents/testingroi.pdf | 04:00 |
| locke105 | i pretty much single-handedly built the Gerrit-Jenkins-OpenStack system we have at my job | 04:00 |
| mordred | locke105: ossum! | 04:00 |
| locke105 | problem is I now spend more time maintaining it then getting to do openstack dev work :( | 04:00 |
| mordred | well, there is that | 04:00 |
| mordred | locke105: you know, you should hang out in #openstack-infra | 04:01 |
| locke105 | thanks for the the link btw | 04:01 |
| locke105 | i'm usually there when i'm at work | 04:01 |
| mordred | locke105: if you're doing things with gerrit and jenkins | 04:01 |
| mordred | cool | 04:01 |
| locke105 | << off-work handle | 04:01 |
| mordred | just making sure | 04:01 |
| clarkb | locke105: eventually you realize all the fun stuff is in making the automation go :) | 04:01 |
| locke105 | mrodden is my at-work handle | 04:01 |
| locke105 | yea i agree | 04:01 |
| locke105 | chef and puppet are the new toy that i have discovered lately | 04:02 |
| clarkb | mordred: you should review https://review.openstack.org/#/c/32455/ | 04:02 |
| clarkb | mordred: if I can get another +2 on that I plan to roll it out tomorrow | 04:02 |
| mordred | locke105: well, all of our infra is in re-consumable puppet | 04:02 |
| clarkb | I expect a fair amount of babysitting will be required | 04:02 |
| clarkb | because firewalls and things... which we unfortunately don't have a good story for automating sanely yet | 04:03 |
| mordred | clarkb: heat! | 04:03 |
| locke105 | mordred: yeah i've looked through the modules, I am more familiar with chef currently and need to study up on puppet more | 04:03 |
| clarkb | mordred: yeah when it works it should be awesome | 04:04 |
| mordred | clarkb: wow. you have a thing which talks gear and zmq in the same app | 04:04 |
| clarkb | mordred: yup :) I will be able to get rid of zmq when we make zuul speak gear | 04:05 |
| mordred | clarkb: neat! | 04:05 |
| clarkb | but until then I have to translate the existing event bus | 04:05 |
| mordred | clarkb: we're about to roll that out aren't we? | 04:05 |
| clarkb | though, I may just keep zmq because this gives me all events not just zuul events | 04:05 |
| clarkb | mordred: yeah I think there are a couple bugs jeblair wants to iron out first | 04:06 |
| *** otherwiseguy has quit IRC | 04:06 | |
| *** kaushikc has joined #openstack-dev | 04:06 | |
| *** dripton_ has joined #openstack-dev | 04:06 | |
| clarkb | mordred: zmq is actually fairly decent and firehosing json at a bunch of subscribers | 04:06 |
| clarkb | s/ant/at/ | 04:06 |
| vishy | smoser: might be this that is needed | 04:06 |
| vishy | smoser: https://lists.openwrt.org/pipermail/openwrt-devel/2011-January/009470.html | 04:06 |
| locke105 | lol @ firehosing | 04:07 |
| mordred | clarkb: yeah. when I look at the oslo.rpc stuff, it actuall makes me want a deployment that uses different tech for the different mesaging types | 04:07 |
| clarkb | mordred: it seems to work well with the logstash stuff. I treat gearman as a job dispatcher and zmq as a firehose | 04:08 |
| clarkb | they are distinct use cases | 04:08 |
| mordred | ++ | 04:08 |
| clarkb | mordred: also newer logstash speaks zmq directly. so I may end up taking advantage of that | 04:09 |
| clarkb | there is a zmq filter which is basically using zmq to dispatch some external filter on an event. gearman probably represents that better | 04:10 |
| *** danwent has joined #openstack-dev | 04:10 | |
| *** dripton has quit IRC | 04:10 | |
| *** carl_baldwin has joined #openstack-dev | 04:13 | |
| mordred | dstufft: when pip is going to install something, and there is an egg file that's been downloaded into the current working directory already (let's say by setuptools) how does it find that? | 04:16 |
| *** nati_ueno has joined #openstack-dev | 04:17 | |
| dstufft | mordred: in certain cases python adds the cwd to the sys.path | 04:18 |
| dstufft | IIRC it basically depends on that | 04:18 |
| dstufft | and it doesn't always work (I've had to add sys.path hacks to make setup_requires stuff get found) | 04:18 |
| dstufft | https://github.com/dstufft/bcrypt/blob/master/setup.py#L22-L24 | 04:19 |
| mordred | dstufft: ah. was afraid of that | 04:19 |
| mordred | dstufft: yeah, I can't figure out rhyme nor reason as to why it can seem to find them sometimes and why othertimes not so much | 04:19 |
| dstufft | I forget the cases python adds sys.path to the cwd | 04:20 |
| dstufft | er | 04:20 |
| dstufft | reverse those | 04:20 |
| dstufft | and I think setuptools might have special support for it | 04:21 |
| mordred | well, adding cwd to sys.path still doesn't make it find eggs in the dir | 04:21 |
| *** lloydde has joined #openstack-dev | 04:21 | |
| dstufft | so it might depend on if it's happning in pip related code, or setuptools related code | 04:21 |
| mordred | grumble | 04:21 |
| mordred | gotcha | 04:21 |
| dstufft | I *don't* have setuptools in my head completely yet | 04:21 |
| mordred | that's probably for the best | 04:22 |
| mordred | now that it's merged and up, I think I might start sending some patches up | 04:22 |
| mordred | like, for the love of god, can we make command objects new style classes so that subclasses can use super()? | 04:22 |
| dstufft | probably not at this point | 04:23 |
| * mordred cries | 04:23 | |
| *** jbresnah has joined #openstack-dev | 04:23 | |
| dstufft | Part of my goals around packaging in the future is getting it the fuck out of stdlib | 04:24 |
| mordred | yes please | 04:25 |
| mordred | that's part of what kills this | 04:25 |
| dstufft | PEP uh | 04:26 |
| dstufft | 439? | 04:26 |
| dstufft | is what enables that | 04:26 |
| mordred | of course, even that - part of my pain is that I'm not going to see changes that are going in to core python for years and years | 04:26 |
| dstufft | http://www.python.org/dev/peps/pep-0439/ | 04:26 |
| dstufft | mordred: well PEP439 is just including pip in python by default | 04:26 |
| dstufft | well a bootstrap script that will get pip | 04:26 |
| mordred | ++ | 04:26 |
| mordred | yeah | 04:26 |
| mordred | you don't want pip in python | 04:27 |
| *** adalbas has quit IRC | 04:27 | |
| dstufft | and part of that work is making pip able to install from wheels without setuptools | 04:27 |
| dstufft | sdists will still require setuptools | 04:27 |
| mordred | that's fine | 04:28 |
| dstufft | but for older pythons the only deficiency will be that you'll need to figure out how to get your own pip | 04:28 |
| mordred | yup. eventually though, we'll stop caring about those pythons | 04:28 |
| dstufft | so older pythons actually benefit a lot from moving stuff out of stdlib, because they can get new work | 04:28 |
| mordred | you konw, 10 years from now | 04:28 |
| mordred | ++ | 04:28 |
| *** basha has joined #openstack-dev | 04:28 | |
| *** nati_ueno has quit IRC | 04:31 | |
| *** vipul is now known as vipul|away | 04:32 | |
| dstufft | mordred: yea we learned backwards compat on the order of decades kinda sucks | 04:33 |
| dstufft | for packaging anyways | 04:33 |
| dstufft | I was 10 y/o when distuils merged into CPython :V | 04:33 |
| locke105 | mordred: totally bookmarked that slide set you linked :) | 04:33 |
| mordred | locke105: sweet! | 04:33 |
| locke105 | definitely worth a read | 04:34 |
| mordred | hrm | 04:37 |
| *** radsy has quit IRC | 04:37 | |
| mordred | dstufft: if I just eat the migration cost for folks of distribute -> setuptools 0.7 as needing to be manual right now | 04:37 |
| mordred | dstufft: I think I might be able to remove some complexity | 04:37 |
| mordred | dstufft: other than the distribute problem, is it reasonably ready to go? | 04:38 |
| dstufft | mordred: I don't think pip has been tested with it yet | 04:39 |
| dstufft | so no promises on compat until 1.4 lands | 04:39 |
| dstufft | well | 04:40 |
| dstufft | the setuptools guy says he confirmed it still works | 04:40 |
| dstufft | but pip devs haven't | 04:40 |
| dstufft | brb reboot | 04:40 |
| *** dstufft has quit IRC | 04:40 | |
| *** souvik has joined #openstack-dev | 04:42 | |
| lifeless | what's moniker's new name? | 04:42 |
| lifeless | ah designate | 04:42 |
| *** SergeyLukjanov has quit IRC | 04:43 | |
| *** prekarat has joined #openstack-dev | 04:44 | |
| *** PhongLy has quit IRC | 04:45 | |
| *** prekarat1 has joined #openstack-dev | 04:45 | |
| *** dstufft has joined #openstack-dev | 04:45 | |
| *** jamielennox|away is now known as jamielennox | 04:46 | |
| dstufft | mordred: fwiw pip 1.4 (and probably a 1.3.x) will support upgrading distribute too | 04:46 |
| dstufft | mordred: no idea what jason did to test pip + setuptools 0.7 | 04:46 |
| mordred | dstufft: well, I think I'm going to take a quick tack at removing the explicit distribute depend from our stuff and seeing what breaks | 04:47 |
| *** prekarat has quit IRC | 04:48 | |
| mordred | because I think I'm doing a lot of work around distribute issues that might could just go away (which would make me happy) | 04:49 |
| *** andrew_plunk has joined #openstack-dev | 04:51 | |
| *** sleepsonthefloor has joined #openstack-dev | 04:51 | |
| dstufft | mordred: as far as the pip maintainers go, we're just starting testing with 0.7, we were hoping we'd release pip, venv, and setuptool 0.7 together all well tested and documented to work together, he kinda forged ahead and released it so now we're scrambling to make sure we're compat :/ | 04:54 |
| mordred | dstufft: stellar | 04:54 |
| dstufft | if you try it please do tell us any issues you come across | 04:55 |
| dstufft | https://github.com/pypa/pip/issues/941 https://github.com/pypa/pip/issues/986 relevant issues | 04:55 |
| *** jbresnah has quit IRC | 04:56 | |
| mordred | dstufft: oh, I shall | 04:56 |
| *** danwent has quit IRC | 04:56 | |
| *** stevemar has quit IRC | 04:58 | |
| *** andrew_plunk has quit IRC | 04:59 | |
| *** carl_baldwin has quit IRC | 05:03 | |
| *** martine_ has quit IRC | 05:03 | |
| openstackgerrit | A change was merged to openstack/cinder: NFS drivers don't honor vm size with volume from an image https://review.openstack.org/30298 | 05:04 |
| openstackgerrit | A change was merged to openstack/nova: Cleanup nova exception message conversion https://review.openstack.org/32074 | 05:04 |
| *** aeperezt has quit IRC | 05:05 | |
| *** rushiagr has joined #openstack-dev | 05:15 | |
| *** koolhead17 has quit IRC | 05:17 | |
| *** mdomsch has joined #openstack-dev | 05:18 | |
| *** matiu has quit IRC | 05:19 | |
| *** avishay has joined #openstack-dev | 05:24 | |
| *** PhongLy has joined #openstack-dev | 05:30 | |
| *** neelashah has quit IRC | 05:30 | |
| *** rushiagr has quit IRC | 05:37 | |
| *** basha has quit IRC | 05:38 | |
| *** stevebaker has joined #openstack-dev | 05:38 | |
| *** stevebaker has left #openstack-dev | 05:39 | |
| *** navid_ has joined #openstack-dev | 05:40 | |
| *** basha has joined #openstack-dev | 05:40 | |
| *** armax has joined #openstack-dev | 05:40 | |
| *** mrunge has joined #openstack-dev | 05:41 | |
| *** Aarti has joined #openstack-dev | 05:43 | |
| *** freedomhui has joined #openstack-dev | 05:46 | |
| *** Aarti has quit IRC | 05:50 | |
| *** Aarti has joined #openstack-dev | 05:51 | |
| *** jbresnah has joined #openstack-dev | 05:53 | |
| *** lonelyibex has joined #openstack-dev | 05:57 | |
| *** lloydde has quit IRC | 05:59 | |
| *** lloydde has joined #openstack-dev | 05:59 | |
| *** PhongLy1 has joined #openstack-dev | 06:00 | |
| *** PhongLy has quit IRC | 06:01 | |
| *** lloydde has quit IRC | 06:04 | |
| *** Mandell has quit IRC | 06:06 | |
| *** basha has quit IRC | 06:07 | |
| *** eglynn has joined #openstack-dev | 06:08 | |
| *** Ryan_Lane has joined #openstack-dev | 06:10 | |
| *** souvik has quit IRC | 06:10 | |
| *** eglynn has quit IRC | 06:11 | |
| *** alop has joined #openstack-dev | 06:12 | |
| *** vartom19 has joined #openstack-dev | 06:12 | |
| *** Ryan_Lane1 has joined #openstack-dev | 06:13 | |
| *** Ryan_Lane has quit IRC | 06:14 | |
| *** lonelyibex has quit IRC | 06:14 | |
| openstackgerrit | A change was merged to openstack/keystone: add ca_key to sample configuration https://review.openstack.org/32260 | 06:15 |
| openstackgerrit | A change was merged to openstack/keystone: Commit transaction in migration https://review.openstack.org/32207 | 06:15 |
| *** basha has joined #openstack-dev | 06:15 | |
| openstackgerrit | A change was merged to openstack/nova: Imported Translations from Transifex https://review.openstack.org/31694 | 06:18 |
| *** adalbas has joined #openstack-dev | 06:22 | |
| *** jcoufal has joined #openstack-dev | 06:22 | |
| *** yolanda has joined #openstack-dev | 06:23 | |
| *** jhesketh_ has quit IRC | 06:24 | |
| *** SergeyLukjanov has joined #openstack-dev | 06:24 | |
| *** basha has quit IRC | 06:28 | |
| *** freedomhui has quit IRC | 06:29 | |
| *** sungju has quit IRC | 06:29 | |
| *** lloydde has joined #openstack-dev | 06:29 | |
| *** stevebaker has joined #openstack-dev | 06:30 | |
| *** jbresnah has quit IRC | 06:30 | |
| *** Ryan_Lane1 has quit IRC | 06:31 | |
| *** anfrolov has quit IRC | 06:32 | |
| *** lloydde has quit IRC | 06:34 | |
| *** sthaha has quit IRC | 06:36 | |
| *** basha has joined #openstack-dev | 06:38 | |
| *** belmoreira has joined #openstack-dev | 06:40 | |
| *** jkordish has quit IRC | 06:40 | |
| *** hartsocks has joined #openstack-dev | 06:40 | |
| *** jkordish has joined #openstack-dev | 06:40 | |
| *** stevebaker has left #openstack-dev | 06:43 | |
| *** basha has quit IRC | 06:44 | |
| *** hemna has quit IRC | 06:44 | |
| *** mrunge has quit IRC | 06:45 | |
| *** Mandell has joined #openstack-dev | 06:46 | |
| *** eglynn has joined #openstack-dev | 06:48 | |
| *** mrunge has joined #openstack-dev | 06:49 | |
| *** rushiagr has joined #openstack-dev | 06:51 | |
| *** xBsd has joined #openstack-dev | 06:54 | |
| *** yolanda has quit IRC | 06:55 | |
| *** alop has quit IRC | 06:56 | |
| *** hemna has joined #openstack-dev | 06:56 | |
| *** eglynn has quit IRC | 06:57 | |
| *** giulivo has joined #openstack-dev | 06:58 | |
| *** yolanda has joined #openstack-dev | 06:59 | |
| *** jhesketh has joined #openstack-dev | 07:00 | |
| *** reidrac has joined #openstack-dev | 07:00 | |
| *** ffio has quit IRC | 07:01 | |
| *** ffio_ has joined #openstack-dev | 07:01 | |
| *** xga has joined #openstack-dev | 07:02 | |
| *** Chrisy_ has quit IRC | 07:03 | |
| *** Chrisy has joined #openstack-dev | 07:03 | |
| *** harlowja has quit IRC | 07:06 | |
| *** jtomasek has joined #openstack-dev | 07:06 | |
| *** aloga has quit IRC | 07:07 | |
| openstackgerrit | A change was merged to openstack/cinder: Replace FLAGS with cfg.CONF in api https://review.openstack.org/31487 | 07:08 |
| *** aloga has joined #openstack-dev | 07:09 | |
| *** avishay has quit IRC | 07:10 | |
| *** corXi has joined #openstack-dev | 07:12 | |
| *** shang has joined #openstack-dev | 07:14 | |
| *** jprovazn has joined #openstack-dev | 07:15 | |
| *** amotoki has joined #openstack-dev | 07:18 | |
| *** devvesa has joined #openstack-dev | 07:19 | |
| lifeless | annegentle: whats the right way to start a new manual ? | 07:19 |
| *** flaper87 has joined #openstack-dev | 07:19 | |
| *** ondergetekende has joined #openstack-dev | 07:19 | |
| lifeless | annegentle: pointers to docs etc appreciated :). Also does a project need to be official [incubated/integrated] before it's appropriate to have such a manual ? | 07:19 |
| *** jbresnah has joined #openstack-dev | 07:23 | |
| *** ndipanov has quit IRC | 07:23 | |
| *** ndipanov has joined #openstack-dev | 07:24 | |
| *** armax has quit IRC | 07:25 | |
| *** terryh has joined #openstack-dev | 07:25 | |
| *** tayyab has joined #openstack-dev | 07:26 | |
| *** kaushikc has quit IRC | 07:26 | |
| *** shardy_afk is now known as shardy | 07:28 | |
| *** anfrolov has joined #openstack-dev | 07:28 | |
| *** lloydde has joined #openstack-dev | 07:30 | |
| *** afazekas_ has joined #openstack-dev | 07:30 | |
| *** Mandell has quit IRC | 07:31 | |
| *** jbresnah has quit IRC | 07:33 | |
| *** lloydde has quit IRC | 07:34 | |
| *** JordanP has joined #openstack-dev | 07:36 | |
| openstackgerrit | A change was merged to openstack/oslo-incubator: Move wsgi module to deprecated package https://review.openstack.org/30972 | 07:38 |
| *** sushils has joined #openstack-dev | 07:40 | |
| *** mmagr has joined #openstack-dev | 07:41 | |
| *** sushils has quit IRC | 07:41 | |
| *** boris-42 has quit IRC | 07:42 | |
| *** terryh has quit IRC | 07:42 | |
| openstackgerrit | A change was merged to openstack/oslo-incubator: MatchMakerStub - make it work & add test https://review.openstack.org/30905 | 07:43 |
| *** sushils has joined #openstack-dev | 07:43 | |
| *** michchap has quit IRC | 07:44 | |
| *** anfrolov has quit IRC | 07:44 | |
| *** michchap has joined #openstack-dev | 07:45 | |
| *** xBsd has quit IRC | 07:45 | |
| *** salv-orlando has joined #openstack-dev | 07:49 | |
| *** jbresnah has joined #openstack-dev | 07:53 | |
| *** fbo has joined #openstack-dev | 07:55 | |
| *** briancline has quit IRC | 08:00 | |
| *** chandankumar has joined #openstack-dev | 08:00 | |
| *** briancline has joined #openstack-dev | 08:00 | |
| *** Mandell has joined #openstack-dev | 08:01 | |
| *** AnilV4 has quit IRC | 08:05 | |
| *** salv-orlando has quit IRC | 08:05 | |
| *** kaushikc has joined #openstack-dev | 08:06 | |
| *** tayyab has quit IRC | 08:06 | |
| *** vishal has quit IRC | 08:07 | |
| *** eglynn has joined #openstack-dev | 08:09 | |
| openstackgerrit | A change was merged to openstack/nova: Add missing tests for s3_image_* methods https://review.openstack.org/31034 | 08:09 |
| *** sungju has joined #openstack-dev | 08:09 | |
| *** kaushikc has quit IRC | 08:16 | |
| *** sungju has quit IRC | 08:16 | |
| *** kaushikc has joined #openstack-dev | 08:16 | |
| *** afazekas has quit IRC | 08:16 | |
| *** Mandell has quit IRC | 08:17 | |
| *** derekh has joined #openstack-dev | 08:17 | |
| *** Mandell has joined #openstack-dev | 08:18 | |
| *** jpich has joined #openstack-dev | 08:18 | |
| *** eglynn has quit IRC | 08:19 | |
| *** avishay has joined #openstack-dev | 08:20 | |
| *** morgank has joined #openstack-dev | 08:21 | |
| *** Alexei_987 has joined #openstack-dev | 08:25 | |
| *** lucasagomes has joined #openstack-dev | 08:25 | |
| *** shaon has joined #openstack-dev | 08:27 | |
| *** johnthetubaguy has joined #openstack-dev | 08:27 | |
| *** romcheg has joined #openstack-dev | 08:29 | |
| *** jongleur has joined #openstack-dev | 08:29 | |
| *** matiu has joined #openstack-dev | 08:29 | |
| *** matiu has joined #openstack-dev | 08:29 | |
| jongleur | Hi. Is this the right place to ask/suggest about devstack? | 08:29 |
| *** mkollaro has joined #openstack-dev | 08:30 | |
| jd__ | probably | 08:30 |
| *** afazekas has joined #openstack-dev | 08:30 | |
| *** lloydde has joined #openstack-dev | 08:30 | |
| jongleur | I installed devstack from git as described on devstack.org, but missed python-netaddr. Probably that should be added somehow somewhere as hint or dependency? (dependency might be impossible as devstack is not packaged somehow...) | 08:31 |
| lifeless | jongleur: file a bug please ;) - bugs.launchpad.net/devstack | 08:31 |
| jongleur | lifeless: okay, thanks | 08:31 |
| *** alexpilotti has joined #openstack-dev | 08:32 | |
| jongleur | okay, it's already filed as #1097667, thanks anyway | 08:33 |
| *** eglynn has joined #openstack-dev | 08:33 | |
| *** sungju has joined #openstack-dev | 08:34 | |
| *** AnilV4 has joined #openstack-dev | 08:34 | |
| *** lloydde has quit IRC | 08:35 | |
| *** sungju has quit IRC | 08:38 | |
| lifeless | jongleur: thanks | 08:39 |
| jongleur | you're welcome | 08:39 |
| *** lukego has joined #openstack-dev | 08:42 | |
| *** boris-42 has joined #openstack-dev | 08:46 | |
| *** boris-42 has quit IRC | 08:46 | |
| *** garyk has joined #openstack-dev | 08:46 | |
| *** boris-42 has joined #openstack-dev | 08:46 | |
| *** pixelbeat has joined #openstack-dev | 08:48 | |
| *** jongleur has quit IRC | 08:48 | |
| *** jongleur has joined #openstack-dev | 08:50 | |
| *** psedlak has joined #openstack-dev | 08:52 | |
| *** xqueralt has joined #openstack-dev | 08:53 | |
| *** jbresnah has quit IRC | 08:53 | |
| *** basha has joined #openstack-dev | 08:55 | |
| *** darjeeling has quit IRC | 08:56 | |
| *** openstackstatus has quit IRC | 09:05 | |
| *** shaon_ has joined #openstack-dev | 09:06 | |
| *** shaon has quit IRC | 09:07 | |
| *** vishal has joined #openstack-dev | 09:11 | |
| *** chandankumar has quit IRC | 09:14 | |
| *** prekarat1 has left #openstack-dev | 09:19 | |
| *** timello has quit IRC | 09:23 | |
| *** timello has joined #openstack-dev | 09:24 | |
| *** ogelbukh1 is now known as ogelbukh | 09:26 | |
| *** freedomhui has joined #openstack-dev | 09:27 | |
| *** lloydde has joined #openstack-dev | 09:31 | |
| *** rushiagr1 has joined #openstack-dev | 09:34 | |
| *** rushiagr has quit IRC | 09:34 | |
| *** lloydde has quit IRC | 09:36 | |
| *** Mandell has quit IRC | 09:42 | |
| *** darjeeling has joined #openstack-dev | 09:43 | |
| *** ilyashakhat_ has joined #openstack-dev | 09:51 | |
| *** ilyashakhat has quit IRC | 09:52 | |
| *** markmc has joined #openstack-dev | 09:56 | |
| *** jruzicka has joined #openstack-dev | 09:57 | |
| *** salv-orlando has joined #openstack-dev | 09:59 | |
| *** yolanda has quit IRC | 10:00 | |
| *** fbo has quit IRC | 10:04 | |
| *** shaon has joined #openstack-dev | 10:05 | |
| *** yolanda has joined #openstack-dev | 10:05 | |
| *** sungju has joined #openstack-dev | 10:07 | |
| *** fbo has joined #openstack-dev | 10:07 | |
| *** rods has joined #openstack-dev | 10:09 | |
| *** shaon_ has quit IRC | 10:09 | |
| *** drewlander has joined #openstack-dev | 10:12 | |
| *** vishal has quit IRC | 10:13 | |
| *** romcheg1 has joined #openstack-dev | 10:17 | |
| *** darjeeling has quit IRC | 10:17 | |
| *** corXi has quit IRC | 10:18 | |
| *** mdenny has quit IRC | 10:21 | |
| *** romcheg has quit IRC | 10:21 | |
| *** portante has quit IRC | 10:26 | |
| *** rushiagr has joined #openstack-dev | 10:28 | |
| *** lloydde has joined #openstack-dev | 10:32 | |
| *** rushiagr1 has quit IRC | 10:32 | |
| *** darjeeling has joined #openstack-dev | 10:36 | |
| openstackgerrit | A change was merged to openstack/nova: Normalize path for finding api_samples dir https://review.openstack.org/31450 | 10:36 |
| *** lloydde has quit IRC | 10:36 | |
| openstackgerrit | A change was merged to openstack/nova: Handle security group quota exceeded gracefully https://review.openstack.org/30095 | 10:36 |
| enikanorov_ | garyk: ping | 10:44 |
| garyk | enikanorov-w_: hi | 10:45 |
| garyk | enikanorov_: hi | 10:49 |
| *** freedomhui has quit IRC | 10:50 | |
| *** PhongLy1 has quit IRC | 10:50 | |
| *** Vivek has joined #openstack-dev | 10:55 | |
| *** derekh has quit IRC | 11:03 | |
| jongleur | is it possible to use devstack to setup a physical compute node (that's not a controller node) and attatch that to the first node installed via devstack before? | 11:06 |
| *** belmoreira has quit IRC | 11:07 | |
| *** derekh has joined #openstack-dev | 11:08 | |
| openstackgerrit | A change was merged to openstack/oslo-incubator: Enable user to configure pool_timeout https://review.openstack.org/31753 | 11:08 |
| openstackgerrit | A change was merged to openstack/oslo.config: python3: More python3 compat updates https://review.openstack.org/32143 | 11:08 |
| *** basha has quit IRC | 11:12 | |
| *** basha has joined #openstack-dev | 11:13 | |
| *** mjfork has joined #openstack-dev | 11:14 | |
| *** romcheg has joined #openstack-dev | 11:18 | |
| *** marun has quit IRC | 11:20 | |
| *** romcheg1 has quit IRC | 11:22 | |
| *** lbragstad has quit IRC | 11:26 | |
| *** dims has joined #openstack-dev | 11:27 | |
| *** krtaylor has quit IRC | 11:28 | |
| *** lbragstad has joined #openstack-dev | 11:28 | |
| *** basha has quit IRC | 11:30 | |
| *** pcm__ has joined #openstack-dev | 11:31 | |
| *** lloydde has joined #openstack-dev | 11:32 | |
| *** basha has joined #openstack-dev | 11:33 | |
| *** johnthetubaguy has quit IRC | 11:34 | |
| *** krtaylor has joined #openstack-dev | 11:34 | |
| *** dhellmann-away has quit IRC | 11:34 | |
| *** darjeeling has quit IRC | 11:35 | |
| *** HenryG has joined #openstack-dev | 11:35 | |
| *** marun has joined #openstack-dev | 11:36 | |
| *** lloydde has quit IRC | 11:37 | |
| *** johnthetubaguy has joined #openstack-dev | 11:38 | |
| *** belmoreira has joined #openstack-dev | 11:38 | |
| *** dhellmann has joined #openstack-dev | 11:39 | |
| *** lbragstad has quit IRC | 11:39 | |
| *** henrynash has joined #openstack-dev | 11:40 | |
| *** belmoreira has quit IRC | 11:47 | |
| *** derekh has quit IRC | 11:53 | |
| *** basha has quit IRC | 11:53 | |
| *** derekh has joined #openstack-dev | 11:54 | |
| giulivo | I think I pasted my question in the wrong chan | 11:54 |
| giulivo | dtroyer_zz or anyone else willing to help with my question on #openstack ? | 11:55 |
| giulivo | for devstack, do we prefer the component's dependencies to be installed via distro packages or rather via pip ? | 11:56 |
| *** SergeyLukjanov has quit IRC | 11:56 | |
| *** basha has joined #openstack-dev | 11:58 | |
| *** SergeyLukjanov has joined #openstack-dev | 11:59 | |
| *** SergeyLukjanov has quit IRC | 11:59 | |
| *** dprince has joined #openstack-dev | 11:59 | |
| *** beagles has joined #openstack-dev | 12:00 | |
| *** sandywalsh has quit IRC | 12:01 | |
| *** tmclaugh[work] has joined #openstack-dev | 12:02 | |
| *** mrunge has quit IRC | 12:03 | |
| *** jamielennox is now known as jamielennox|away | 12:05 | |
| *** belmoreira has joined #openstack-dev | 12:05 | |
| *** novas0x2a|laptop has quit IRC | 12:05 | |
| *** iartarisi has joined #openstack-dev | 12:05 | |
| *** lucasagomes is now known as lucas-hungry | 12:06 | |
| *** READ10 has quit IRC | 12:11 | |
| *** egallen has joined #openstack-dev | 12:13 | |
| *** martine_ has joined #openstack-dev | 12:14 | |
| *** sandywalsh has joined #openstack-dev | 12:14 | |
| *** morazi has quit IRC | 12:15 | |
| *** zyluo has joined #openstack-dev | 12:15 | |
| *** sthaha has joined #openstack-dev | 12:16 | |
| *** dims has quit IRC | 12:16 | |
| *** dims has joined #openstack-dev | 12:17 | |
| *** katylava has joined #openstack-dev | 12:23 | |
| *** kaushikc has quit IRC | 12:25 | |
| *** kaushikc has joined #openstack-dev | 12:25 | |
| *** jergerber has joined #openstack-dev | 12:26 | |
| *** lbragstad has joined #openstack-dev | 12:29 | |
| *** basha has quit IRC | 12:30 | |
| *** SergeyLukjanov has joined #openstack-dev | 12:32 | |
| *** lloydde has joined #openstack-dev | 12:33 | |
| *** basha has joined #openstack-dev | 12:33 | |
| *** marun has quit IRC | 12:36 | |
| *** lloydde has quit IRC | 12:37 | |
| *** lucas-hungry is now known as lucasagomes | 12:44 | |
| *** basha has quit IRC | 12:45 | |
| *** darjeeling has joined #openstack-dev | 12:46 | |
| *** redbeard2 has quit IRC | 12:47 | |
| *** matiu has quit IRC | 12:49 | |
| *** avishay has quit IRC | 12:51 | |
| *** michchap has quit IRC | 12:53 | |
| *** michchap_ has joined #openstack-dev | 12:53 | |
| *** michchap_ has quit IRC | 12:53 | |
| *** jasondotstar has joined #openstack-dev | 12:54 | |
| *** Aarti has quit IRC | 12:54 | |
| *** basha has joined #openstack-dev | 12:55 | |
| *** basha has quit IRC | 12:55 | |
| *** erkules_ is now known as erkules | 12:56 | |
| annegentle | Doc team meeting in 3 minutes in #openstack-meeting | 12:57 |
| *** marun has joined #openstack-dev | 12:58 | |
| *** kbringard has joined #openstack-dev | 13:00 | |
| *** pfreund has joined #openstack-dev | 13:00 | |
| *** terryh has joined #openstack-dev | 13:00 | |
| *** blues-man has joined #openstack-dev | 13:00 | |
| *** gargya has joined #openstack-dev | 13:01 | |
| *** lukego has quit IRC | 13:02 | |
| *** dkehn has quit IRC | 13:03 | |
| *** changbl has quit IRC | 13:04 | |
| *** avishay has joined #openstack-dev | 13:05 | |
| *** dkehn has joined #openstack-dev | 13:06 | |
| *** andrewbogott_afk is now known as andrewbogott | 13:06 | |
| *** belmoreira1 has joined #openstack-dev | 13:08 | |
| *** basha has joined #openstack-dev | 13:08 | |
| *** belmoreira has quit IRC | 13:10 | |
| *** belmoreira1 has quit IRC | 13:11 | |
| *** kaushikc has quit IRC | 13:11 | |
| *** belmoreira has joined #openstack-dev | 13:11 | |
| *** kaushikc has joined #openstack-dev | 13:12 | |
| *** Tross has quit IRC | 13:13 | |
| *** spzala has joined #openstack-dev | 13:13 | |
| agorodnev | hi, I have a question to anybody from oslo team. Is it possible to update some particular file using update.py script? | 13:14 |
| *** prad_ has joined #openstack-dev | 13:15 | |
| *** pabelanger_ has joined #openstack-dev | 13:20 | |
| *** johnthetubaguy has quit IRC | 13:20 | |
| *** pabelanger_ has quit IRC | 13:21 | |
| *** pabelanger_ has joined #openstack-dev | 13:21 | |
| *** pabelanger has quit IRC | 13:21 | |
| *** garyk has quit IRC | 13:21 | |
| *** pabelanger_ is now known as pabelanger | 13:22 | |
| *** markmcclain has joined #openstack-dev | 13:22 | |
| *** tzumainn has joined #openstack-dev | 13:22 | |
| *** pabelanger_ has joined #openstack-dev | 13:22 | |
| *** markmcclain has quit IRC | 13:22 | |
| *** markmcclain has joined #openstack-dev | 13:22 | |
| *** johnthetubaguy has joined #openstack-dev | 13:23 | |
| *** garyk has joined #openstack-dev | 13:23 | |
| *** davidkranz has joined #openstack-dev | 13:23 | |
| *** michchap has joined #openstack-dev | 13:24 | |
| *** vartom19 has quit IRC | 13:24 | |
| *** morazi has joined #openstack-dev | 13:24 | |
| *** gakott has joined #openstack-dev | 13:26 | |
| *** gargya_ has joined #openstack-dev | 13:27 | |
| *** rharwood has joined #openstack-dev | 13:27 | |
| *** neelashah has joined #openstack-dev | 13:27 | |
| *** markmcclain has quit IRC | 13:27 | |
| *** gargya has quit IRC | 13:28 | |
| *** bknudson has joined #openstack-dev | 13:28 | |
| *** gargya has joined #openstack-dev | 13:29 | |
| *** krtaylor has quit IRC | 13:29 | |
| *** skort has joined #openstack-dev | 13:29 | |
| *** garyk has quit IRC | 13:30 | |
| *** jayg|g0n3 is now known as jayg | 13:31 | |
| *** gargya_ has quit IRC | 13:32 | |
| *** gakott has quit IRC | 13:33 | |
| *** abhishekkr has quit IRC | 13:33 | |
| *** lloydde has joined #openstack-dev | 13:33 | |
| *** michchap has quit IRC | 13:33 | |
| *** litong has joined #openstack-dev | 13:34 | |
| *** timello_ has joined #openstack-dev | 13:36 | |
| *** pmathews has joined #openstack-dev | 13:36 | |
| *** afazekas_ has quit IRC | 13:37 | |
| *** lloydde has quit IRC | 13:38 | |
| *** skort has quit IRC | 13:39 | |
| *** timello has quit IRC | 13:39 | |
| *** skort has joined #openstack-dev | 13:39 | |
| *** adalbas has quit IRC | 13:39 | |
| *** jecarey has joined #openstack-dev | 13:42 | |
| *** rnirmal has joined #openstack-dev | 13:42 | |
| *** johnthetubaguy has quit IRC | 13:45 | |
| *** johnthetubaguy has joined #openstack-dev | 13:45 | |
| *** abhishekkr has joined #openstack-dev | 13:46 | |
| *** mtreinish has joined #openstack-dev | 13:51 | |
| *** abhishekkr has quit IRC | 13:51 | |
| *** drewlander has quit IRC | 13:52 | |
| *** eharney has joined #openstack-dev | 13:52 | |
| *** jimfehlig has joined #openstack-dev | 13:54 | |
| *** gakott has joined #openstack-dev | 13:54 | |
| *** garyk has joined #openstack-dev | 13:56 | |
| *** PhongLy has joined #openstack-dev | 13:56 | |
| *** ladquin is now known as ladquin_afk | 13:56 | |
| *** stevemar has joined #openstack-dev | 13:57 | |
| *** JordanP has quit IRC | 13:57 | |
| *** garyTh has joined #openstack-dev | 13:57 | |
| *** skort has quit IRC | 13:58 | |
| *** drewlander has joined #openstack-dev | 13:59 | |
| *** Ryan_Lane has joined #openstack-dev | 13:59 | |
| *** gakott has quit IRC | 13:59 | |
| *** michchap has joined #openstack-dev | 14:00 | |
| *** dspano has joined #openstack-dev | 14:00 | |
| *** networkstatic has quit IRC | 14:01 | |
| *** networkstatic has joined #openstack-dev | 14:01 | |
| *** michchap_ has joined #openstack-dev | 14:02 | |
| *** souvik has joined #openstack-dev | 14:02 | |
| *** changbl has joined #openstack-dev | 14:03 | |
| *** michcha__ has joined #openstack-dev | 14:03 | |
| *** michchap has quit IRC | 14:04 | |
| *** michchap has joined #openstack-dev | 14:05 | |
| *** michch___ has joined #openstack-dev | 14:06 | |
| *** michchap_ has quit IRC | 14:07 | |
| *** michchap_ has joined #openstack-dev | 14:07 | |
| *** timello_ has quit IRC | 14:07 | |
| *** michcha__ has quit IRC | 14:08 | |
| *** rcleere has quit IRC | 14:08 | |
| *** michcha__ has joined #openstack-dev | 14:09 | |
| *** henrynash has quit IRC | 14:09 | |
| *** michchap has quit IRC | 14:09 | |
| *** amotoki has quit IRC | 14:09 | |
| *** chuckieb has joined #openstack-dev | 14:10 | |
| *** timello_ has joined #openstack-dev | 14:10 | |
| *** michch___ has quit IRC | 14:10 | |
| *** cp16net is now known as cp16net|away | 14:11 | |
| *** michchap_ has quit IRC | 14:11 | |
| *** adalbas has joined #openstack-dev | 14:11 | |
| *** sld is now known as sld-away | 14:13 | |
| *** msmedved has joined #openstack-dev | 14:13 | |
| *** michcha__ has quit IRC | 14:14 | |
| *** basha has quit IRC | 14:15 | |
| *** cp16net|away is now known as cp16net | 14:16 | |
| *** cp16net is now known as cp16net|away | 14:16 | |
| *** READ10 has joined #openstack-dev | 14:16 | |
| *** roz has quit IRC | 14:17 | |
| *** roz has joined #openstack-dev | 14:17 | |
| *** cp16net|away is now known as cp16net | 14:18 | |
| *** marun has quit IRC | 14:18 | |
| *** JordanP has joined #openstack-dev | 14:20 | |
| *** basha has joined #openstack-dev | 14:21 | |
| *** nunosantos has joined #openstack-dev | 14:21 | |
| *** noslzzp has joined #openstack-dev | 14:22 | |
| *** Ruetobas has quit IRC | 14:23 | |
| *** basha has quit IRC | 14:23 | |
| *** Ruetobas has joined #openstack-dev | 14:24 | |
| openstackgerrit | A change was merged to openstack/horizon: Edit some load balancers titles https://review.openstack.org/32393 | 14:24 |
| *** Ruetobas has quit IRC | 14:26 | |
| *** krtaylor has joined #openstack-dev | 14:26 | |
| *** datsun180b has joined #openstack-dev | 14:27 | |
| *** freedomhui has joined #openstack-dev | 14:27 | |
| *** enikanorov_ has quit IRC | 14:28 | |
| *** sleepsonthefloor has quit IRC | 14:31 | |
| *** esheffield has joined #openstack-dev | 14:31 | |
| *** Ruetobas has joined #openstack-dev | 14:32 | |
| *** kaushikc has quit IRC | 14:32 | |
| *** dkehn has quit IRC | 14:33 | |
| *** marun has joined #openstack-dev | 14:37 | |
| *** rcleere has joined #openstack-dev | 14:37 | |
| *** FunnyLookinHat has joined #openstack-dev | 14:39 | |
| *** galstrom_zzz is now known as galstrom | 14:39 | |
| *** ondergetekende has quit IRC | 14:40 | |
| *** portante has joined #openstack-dev | 14:40 | |
| *** otherwiseguy has joined #openstack-dev | 14:40 | |
| *** jvrbanac has joined #openstack-dev | 14:40 | |
| *** topol has joined #openstack-dev | 14:40 | |
| *** michchap has joined #openstack-dev | 14:40 | |
| *** Ryan_Lane has quit IRC | 14:43 | |
| *** redbeard2 has joined #openstack-dev | 14:43 | |
| *** aeperezt has joined #openstack-dev | 14:43 | |
| *** monst has joined #openstack-dev | 14:44 | |
| *** jcoufal has quit IRC | 14:44 | |
| *** mdenny has joined #openstack-dev | 14:44 | |
| *** lloydde has joined #openstack-dev | 14:47 | |
| *** bashok has joined #openstack-dev | 14:48 | |
| *** jtomasek has quit IRC | 14:48 | |
| *** alexpilotti has quit IRC | 14:49 | |
| *** kevinconway has joined #openstack-dev | 14:50 | |
| *** BalleS_ has quit IRC | 14:52 | |
| *** andrew_plunk has joined #openstack-dev | 14:52 | |
| *** michchap has quit IRC | 14:52 | |
| *** michchap has joined #openstack-dev | 14:54 | |
| *** n0ano has joined #openstack-dev | 14:54 | |
| *** corXi has joined #openstack-dev | 14:55 | |
| mordred | markmc: morning | 14:55 |
| *** armax has joined #openstack-dev | 14:55 | |
| markmc | mordred, morning of ambiguous qualities? :) | 14:55 |
| markmc | mordred, do we just go ahead and tag 0.5.15 now and give it a shot? | 14:56 |
| mordred | markmc: I mean, there is a morning and it has qualities : | 14:56 |
| *** jongleur has quit IRC | 14:56 | |
| mordred | markmc: I was just about to ask that question myself. do you think just the pip change is still ok for another 0.5 change instead of a 0.6 bump? | 14:56 |
| *** michchap_ has joined #openstack-dev | 14:57 | |
| *** susanne-balle has joined #openstack-dev | 14:57 | |
| markmc | mordred, well, since we've capped everything at <0.6 I assumed the plan was to only bump to 0.6 if we made a compat breaking change | 14:57 |
| markmc | mordred, this should be compatible | 14:57 |
| markmc | mordred, bugs aside | 14:58 |
| *** souvik has quit IRC | 14:58 | |
| *** sushils has quit IRC | 14:58 | |
| mordred | markmc: indeed. just making sure you agreed there | 14:58 |
| markmc | mordred, cool | 14:58 |
| *** michcha__ has joined #openstack-dev | 14:58 | |
| *** souvik has joined #openstack-dev | 14:58 | |
| *** michchap has quit IRC | 14:58 | |
| mordred | markmc: also, re the test patch - last I tried to follow up with lifeless on his comment I got nowhere, so I was thinking of it like this - currently the testr class exists and works fine | 14:59 |
| *** michchap has joined #openstack-dev | 14:59 | |
| mordred | markmc: a possible step forward is to move the testr setuptools class into pbr and remove it from testr - but that can be done without breaking pbr functionality | 14:59 |
| mordred | or it might be that we make the testr class a class that testr intends to export | 15:00 |
| markmc | mordred, I'd have preferred if we were on 1.x.y now and capped everything at <2.0 ... but same difference, I guess | 15:00 |
| mordred | markmc: I was actually _just_ thinking that myself | 15:00 |
| *** souvik has quit IRC | 15:00 | |
| mordred | markmc: how about when we bump, we bump to 1.x series so we can get on that train properly | 15:00 |
| *** michch___ has joined #openstack-dev | 15:00 | |
| *** danwent has joined #openstack-dev | 15:01 | |
| *** sushils has joined #openstack-dev | 15:01 | |
| markmc | mordred, we'll have pbr version out there that say they work with a certain range of testrepository | 15:01 |
| *** armax has left #openstack-dev | 15:01 | |
| *** michchap_ has quit IRC | 15:01 | |
| markmc | mordred, but what lifeless is saying is he can't guarantee a future version in that range won't break us | 15:01 |
| mordred | GOTCHA | 15:01 |
| mordred | k | 15:02 |
| *** michchap_ has joined #openstack-dev | 15:02 | |
| mordred | I'll re-address | 15:02 |
| *** ffio_ has quit IRC | 15:02 | |
| markmc | mordred, coolness | 15:02 |
| *** _ffio_ has joined #openstack-dev | 15:02 | |
| *** michcha__ has quit IRC | 15:02 | |
| *** jprovazn has quit IRC | 15:02 | |
| mordred | markmc, jd__: 0.5.15 tag pushed | 15:03 |
| mordred | dhellmann: ^^ | 15:03 |
| *** michcha__ has joined #openstack-dev | 15:03 | |
| derekh | talking about pbr?, is there currently a problem with it? | 15:03 |
| derekh | this command used to work "sudo pip install -U git+https://github.com/openstack-dev/pbr.git" | 15:03 |
| * jd__ runs around in this office screaming | 15:03 | |
| derekh | no I get "NameError: name 'egg_info' is not defined" | 15:03 |
| markmc | derekh, heh, nice timing | 15:03 |
| markmc | 2 seconds after we tag a new release | 15:04 |
| markmc | derekh, when did that start happening | 15:04 |
| *** michchap has quit IRC | 15:04 | |
| mordred | derekh: I never expect sudo pip install -U git+https://... to work for any of our projects | 15:04 |
| derekh | markmc: about an hour ago http://54.228.118.193/toci/toci_logs_FAqRJSe/ | 15:04 |
| mordred | oh, wait, no, it should work. ignore | 15:04 |
| mordred | me | 15:04 |
| markmc | derekh, and first step - file a bug | 15:04 |
| mordred | (no coffee for me yet) | 15:04 |
| derekh | markmc: sorry about timing, see full log here http://54.228.118.193/toci/toci_logs_8NlRpBm/setup.out | 15:05 |
| *** reidrac has quit IRC | 15:05 | |
| derekh | markmc: will log a bug | 15:05 |
| *** michch___ has quit IRC | 15:05 | |
| markmc | derekh, timing isn't your fault :) | 15:05 |
| markmc | derekh, thanks for the report | 15:05 |
| derekh | np | 15:05 |
| *** gargya has quit IRC | 15:05 | |
| dstufft | egg_info not being defined sounds like setuptools/distribute got wrecked | 15:06 |
| dstufft | since that's not part of distutils IIRC, only the extensions | 15:06 |
| * mordred agrees with dstufft | 15:06 | |
| *** michchap_ has quit IRC | 15:06 | |
| dave_mcn | Looking for an approval for https://review.openstack.org/#/c/30270/ anyone about who can take a look? | 15:06 |
| mordred | that sounds like the "something tried to install distribute overtop of too-old setuptools and things died" | 15:06 |
| dstufft | oh | 15:07 |
| dstufft | this is convient timing | 15:07 |
| *** mmagr has quit IRC | 15:07 | |
| *** michcha__ has quit IRC | 15:07 | |
| dstufft | looks like we're having the same problem trying to run egg_info twice w/ distribute 0.7 | 15:07 |
| dstufft | where we're == pip devs | 15:07 |
| *** marun has quit IRC | 15:08 | |
| *** mmagr has joined #openstack-dev | 15:09 | |
| *** freedomhui has quit IRC | 15:09 | |
| derekh | markmc: mordred https://bugs.launchpad.net/pbr/+bug/1189941 | 15:10 |
| uvirtbot | Launchpad bug 1189941 in pbr "installing with pip from git no longer works" [Undecided,New] | 15:10 |
| mordred | derekh: thanks | 15:10 |
| *** romcheg has quit IRC | 15:10 | |
| mordred | dstufft: awesom1 | 15:10 |
| *** JordanP has quit IRC | 15:11 | |
| *** garyk has quit IRC | 15:11 | |
| *** garyk has joined #openstack-dev | 15:11 | |
| *** xqueralt is now known as xqueralt-afk | 15:12 | |
| markmc | mordred, could you make oslo-bugs the bug supervisor for pbr and hacking ? | 15:12 |
| dstufft | mordred: fwiw directly depending on either setuptools or distribute tends to be recipe for pain everytime I've done it :( (e.g. in the package metadata) | 15:12 |
| *** xqueralt-afk is now known as xqueralt | 15:12 | |
| *** souvik has joined #openstack-dev | 15:12 | |
| mordred | dstufft: yes. I'm actually starting to work through patches to remove that direct dependency | 15:12 |
| *** egallen has quit IRC | 15:13 | |
| *** davidkranz has quit IRC | 15:16 | |
| *** souvik has quit IRC | 15:16 | |
| *** sandywalsh has quit IRC | 15:16 | |
| *** carl_baldwin has joined #openstack-dev | 15:17 | |
| *** rushiagr has quit IRC | 15:18 | |
| *** davidkranz has joined #openstack-dev | 15:18 | |
| *** jhesketh has quit IRC | 15:18 | |
| mordred | markmc: yes. in the mean time: https://review.openstack.org/32527 - updating that patch solve this, and also makes dstufft happier | 15:19 |
| *** egallen has joined #openstack-dev | 15:19 | |
| mordred | markmc: also, I think we've identified over the last day several new things that we need to test on pbr changes | 15:19 |
| openstackgerrit | A change was merged to openstack/glance: Sql query optimization for image detail https://review.openstack.org/32413 | 15:19 |
| openstackgerrit | A change was merged to openstack/quantum: Imported Translations from Transifex https://review.openstack.org/32559 | 15:19 |
| openstackgerrit | A change was merged to openstack/quantum: Protect PoolStats table from negative values. https://review.openstack.org/32545 | 15:19 |
| dstufft | There needs to be bigger numbers to express how much + I am to that change | 15:19 |
| mordred | markmc: the "install trunk pbr in devstack" does not seem to have been enough | 15:19 |
| markmc | mordred, right | 15:20 |
| *** koolhead17 has joined #openstack-dev | 15:20 | |
| *** derekh has quit IRC | 15:20 | |
| *** mmagr has quit IRC | 15:21 | |
| *** derekh has joined #openstack-dev | 15:21 | |
| *** souvik has joined #openstack-dev | 15:22 | |
| *** yidclare has joined #openstack-dev | 15:23 | |
| *** yjiang5 has joined #openstack-dev | 15:23 | |
| *** lloydde_ has joined #openstack-dev | 15:23 | |
| *** hartsocks1 has joined #openstack-dev | 15:24 | |
| *** hartsocks1 has quit IRC | 15:24 | |
| *** kenperkins_ is now known as kenperkins | 15:24 | |
| *** hartsocks1 has joined #openstack-dev | 15:24 | |
| *** portante has quit IRC | 15:25 | |
| *** hartsocks has quit IRC | 15:26 | |
| *** lloydde has quit IRC | 15:26 | |
| *** davidkranz has quit IRC | 15:27 | |
| *** boris-42 has quit IRC | 15:28 | |
| *** romcheg has joined #openstack-dev | 15:28 | |
| *** davidkranz has joined #openstack-dev | 15:28 | |
| *** aelkikhia has joined #openstack-dev | 15:28 | |
| *** hartsocks1 has quit IRC | 15:29 | |
| *** lloydde_ has quit IRC | 15:30 | |
| mordred | dhellmann: have been discussion distribute/setuptools depends with dstufft in this channel | 15:31 |
| *** hartsocks has joined #openstack-dev | 15:31 | |
| dstufft | I think dhellmann has been discussing it on Twitter with me :V | 15:31 |
| dstufft | well yesterday | 15:31 |
| mordred | hehe | 15:31 |
| * mordred goes to look for that twitter ... | 15:31 | |
| *** changbl has quit IRC | 15:32 | |
| *** sandywalsh has joined #openstack-dev | 15:32 | |
| *** bashok has quit IRC | 15:32 | |
| *** mdomsch has quit IRC | 15:34 | |
| *** wenjianhn has quit IRC | 15:34 | |
| *** mmagr has joined #openstack-dev | 15:34 | |
| *** pabelanger has quit IRC | 15:34 | |
| *** michchap has joined #openstack-dev | 15:35 | |
| *** pabelanger has joined #openstack-dev | 15:35 | |
| *** sandywalsh has quit IRC | 15:35 | |
| *** navid_ has left #openstack-dev | 15:36 | |
| *** romcheg has quit IRC | 15:36 | |
| *** rferrari has joined #openstack-dev | 15:36 | |
| *** JordanP has joined #openstack-dev | 15:37 | |
| *** corXi has quit IRC | 15:37 | |
| *** JordanP has quit IRC | 15:37 | |
| *** JordanP has joined #openstack-dev | 15:38 | |
| *** dripton_ is now known as dripton | 15:38 | |
| *** morgank has quit IRC | 15:39 | |
| dstufft | mordred: lol | 15:40 |
| dstufft | https://github.com/pypa/pip/issues/986#issuecomment-19269921 | 15:40 |
| *** garyk has quit IRC | 15:40 | |
| mordred | dstufft: haha. nice | 15:40 |
| *** hemna has quit IRC | 15:40 | |
| *** hemna has joined #openstack-dev | 15:41 | |
| dhellmann | ah, man, I'm sorry dstufft : I saw your full name on twitter, but didn't make the connection with irc | 15:41 |
| dstufft | dhellmann: no worries :) | 15:41 |
| dstufft | I'm dstufft just about everywhere | 15:41 |
| * dhellmann undestands Alex_Gaynor's comment better now | 15:41 | |
| dstufft | first initial + last name supremacy | 15:41 |
| dhellmann | yeah, at least you use your *name* | 15:42 |
| mordred | markmc: https://review.openstack.org/32527 seems to fix everything for folks | 15:42 |
| * dhellmann looks askance at most of #openstack-infra | 15:42 | |
| mordred | hey man. my last name is TOTALLY ordred | 15:42 |
| dhellmann | lol | 15:42 |
| *** drewlander has quit IRC | 15:42 | |
| markmc | mordred, approved | 15:43 |
| mordred | markmc: awesome. | 15:44 |
| *** feleouet has quit IRC | 15:44 | |
| markmc | mordred, fingers crossed | 15:44 |
| *** savid has joined #openstack-dev | 15:44 | |
| mordred | markmc: and seriously. no new patches to pbr land after this until we get that above into a test | 15:44 |
| markmc | mordred, I'm cool with that :) | 15:45 |
| *** ociuhandu has joined #openstack-dev | 15:45 | |
| *** schwicht has joined #openstack-dev | 15:45 | |
| *** michchap has quit IRC | 15:46 | |
| *** belmoreira has quit IRC | 15:47 | |
| mordred | jd__: ^^ new fix going in. when it lands, I'll tag and release and maybe we can stop having this conversation every day | 15:47 |
| *** sride has joined #openstack-dev | 15:49 | |
| jd__ | haha | 15:49 |
| jd__ | I hope so! :) | 15:49 |
| * dhellmann would also like to talk about something else | 15:50 | |
| *** sandywalsh has joined #openstack-dev | 15:50 | |
| *** belmoreira has joined #openstack-dev | 15:51 | |
| mordred | dhellmann: how about twisted vs. eventlet? | 15:51 |
| *** jvrbanac has quit IRC | 15:52 | |
| *** xqueralt is now known as xqueralt-afk | 15:53 | |
| * dhellmann sighs | 15:54 | |
| *** pnavarro has joined #openstack-dev | 15:55 | |
| * flaper87 cries | 15:55 | |
| *** jcoufal has joined #openstack-dev | 15:55 | |
| *** jcoufal has quit IRC | 15:56 | |
| *** JordanP has quit IRC | 15:56 | |
| *** jcoufal has joined #openstack-dev | 15:56 | |
| dhellmann | mordred, jd__: how about whether to pin msgpack or update kombu? | 15:56 |
| jd__ | dhellmann: can we update kombu? | 15:57 |
| dhellmann | jd__: I'd love to, but pinning msgpack seemed like a quicker fix | 15:57 |
| dhellmann | both? | 15:57 |
| dhellmann | msgpack first, then testing with new kombu? | 15:57 |
| jd__ | I think just adding the right >= to kombu will be enough actually | 15:57 |
| dhellmann | ok | 15:57 |
| dhellmann | I'm not sure how we ended up with the wrong requirement to begin with, I thought that wasn't supposed to be possible any more | 15:58 |
| jd__ | do we know what version we should target? | 15:58 |
| *** openstackstatus has joined #openstack-dev | 15:58 | |
| dhellmann | jd__: https://review.openstack.org/#/c/32022/2/tools/pip-requires | 15:58 |
| *** belmoreira has quit IRC | 15:58 | |
| jd__ | dhellmann: it's not our fault I think it's a problem in kombu's dependencies | 15:58 |
| dhellmann | I thought that was right | 15:58 |
| dhellmann | ah | 15:58 |
| *** JordanP has joined #openstack-dev | 15:58 | |
| jd__ | msgpack breaks compat and old kombu don't know about it and doesn't have a <0.2 | 15:59 |
| dhellmann | well I wonder if that means we should be pinning kombu, too | 15:59 |
| *** mordred has quit IRC | 15:59 | |
| *** mordred has joined #openstack-dev | 15:59 | |
| *** lloydde has joined #openstack-dev | 15:59 | |
| dhellmann | our kombu rule is >=1.0.4 | 15:59 |
| dhellmann | I wonder if that needs an upperbound | 15:59 |
| dhellmann | by "our" I mean the requirements project | 16:00 |
| openstackstatus | NOTICE: pbr change broke the gate, a fix is forthcoming | 16:00 |
| *** reed has joined #openstack-dev | 16:00 | |
| *** ChanServ changes topic to "pbr change broke the gate, a fix is forthcoming" | 16:00 | |
| jd__ | dhellmann: haha | 16:00 |
| jd__ | yeah | 16:00 |
| *** mrodden has quit IRC | 16:00 | |
| dhellmann | when I can run devstack again, I'll see if I can figure out which versions actually work together | 16:00 |
| jd__ | I know 2.1.8 breaks with msgpack >= 0.2 and that kombu >= 2.5.10 is ok | 16:00 |
| dhellmann | ok | 16:01 |
| *** xqueralt-afk is now known as xqueralt | 16:01 | |
| jd__ | so we need something between this two | 16:01 |
| *** koolhead17 has quit IRC | 16:01 | |
| *** mrodden1 has joined #openstack-dev | 16:01 | |
| *** yolanda_ has joined #openstack-dev | 16:01 | |
| *** fbo has quit IRC | 16:01 | |
| *** alexpilotti has joined #openstack-dev | 16:02 | |
| jd__ | actually I can find that, let me dig this | 16:02 |
| dhellmann | well, why not just say kombo>=2.5.10,<2.6 and then set the right version of msgpack for that? | 16:02 |
| *** Alexei_987 has quit IRC | 16:02 | |
| *** alop has joined #openstack-dev | 16:03 | |
| *** souvik1 has joined #openstack-dev | 16:03 | |
| dhellmann | I guess we could do kombu>=1.0.4<2.1.8 and msgpack>=0.1<0.2 to get the older versions | 16:03 |
| jd__ | well basically we have no reason to pin msgpack, we support any version; the fact that kombu fails for some version is our only problem | 16:03 |
| dhellmann | I see | 16:03 |
| *** alop has quit IRC | 16:04 | |
| *** seanrob has joined #openstack-dev | 16:04 | |
| jd__ | basically I want to be able to use msgpack >= 0.2, it's already packaged and distributed | 16:04 |
| *** negronjl` has joined #openstack-dev | 16:04 | |
| dhellmann | well, if they aren't handling their own dependencies properly in the older versions, shouldn't we do it? | 16:04 |
| dhellmann | right | 16:04 |
| dhellmann | so I'm suggesting 2 changes | 16:04 |
| jd__ | I don't think we can handle 2 cases | 16:04 |
| *** colinmcnamara has joined #openstack-dev | 16:04 | |
| dhellmann | first, pin msgpack low so we can get things working now | 16:04 |
| dhellmann | second, pin kombu high and unpin msgpack -- but that would require *way* more testing | 16:05 |
| *** sushil_ has joined #openstack-dev | 16:05 | |
| jd__ | we can't do "msgpack&&kombu>=2.5.10 || msgpack<0.2&&kombu<2.5.10" | 16:05 |
| *** spzala_ has joined #openstack-dev | 16:05 | |
| dhellmann | no, these would be 2 changesets | 16:05 |
| *** Slidey_ has joined #openstack-dev | 16:05 | |
| *** monst_ has joined #openstack-dev | 16:05 | |
| jd__ | commit 5c92e93c9bfd33259d935e30db2c09b891ab37f8 | 16:05 |
| jd__ | Author: Ask Solem <ask@celeryproject.org> | 16:05 |
| jd__ | Date: Mon Nov 12 17:04:40 2012 +0000 | 16:05 |
| *** mmagr has quit IRC | 16:05 | |
| jd__ | 16:05 | |
| jd__ | Now depends on msgpack > 0.2.0 | 16:05 |
| jd__ | 16:05 | |
| jd__ | % git describe --tags 5c92e93c9bfd33259d935e30db2c09b891ab37f8 | 16:06 |
| jd__ | v2.4.7-97-g5c92e93 | 16:06 |
| jd__ | 16:06 | |
| *** shang_ has joined #openstack-dev | 16:06 | |
| dstufft | clarkb: I think you suggested I follow this pbr thing because you hate my inbox, this gerrit thing loves email | 16:06 |
| dhellmann | is that a kombu commit? | 16:06 |
| jd__ | so we need kombu > 2.4.7 | 16:06 |
| jd__ | dhellmann: yes | 16:06 |
| jd__ | dhellmann: how does that sound to you? | 16:06 |
| *** ekarlso has joined #openstack-dev | 16:06 | |
| *** shang has quit IRC | 16:06 | |
| *** shang_ has quit IRC | 16:06 | |
| dhellmann | jd__: I just don't have time to test the latest kombu in all of openstack. :-/ | 16:07 |
| *** shang has joined #openstack-dev | 16:07 | |
| jd__ | dhellmann: the latest kombu is already what's pulled in CI, no? | 16:07 |
| dhellmann | it was not on my devstack installation | 16:07 |
| jd__ | we don't have a higher bound on it | 16:07 |
| jd__ | so I don't see how it would change anything | 16:07 |
| *** iartarisi has quit IRC | 16:08 | |
| jd__ | we're already running CI with msgpack 2.0 and kombu 2.5 | 16:08 |
| * dhellmann is checking what he has installed | 16:08 | |
| *** MIDENN_ has joined #openstack-dev | 16:09 | |
| *** pnavarro has quit IRC | 16:09 | |
| *** krtaylor_ has joined #openstack-dev | 16:09 | |
| *** pnavarro has joined #openstack-dev | 16:10 | |
| mordred | markmc: btw - pip 1.4 is going to ignore installing pre-release things by default | 16:10 |
| dhellmann | jd__: devstack installs kombu 1.4.3 from a system package; tox installs 2.5.10 via pip | 16:10 |
| mordred | markmc: so we'll have a possibly richer story for what to do with pre-release oslo things | 16:10 |
| *** monst_ has quit IRC | 16:10 | |
| markmc | mordred, hmm, ok | 16:10 |
| mordred | markmc: in that we can upload them to pypi for ease of depending on them, but you won't get them installed unless you explicitly request their version | 16:10 |
| vishy | smoser: did you see my commentary yesterday? | 16:11 |
| *** negronjl has quit IRC | 16:11 | |
| *** souvik has quit IRC | 16:11 | |
| *** Slidey has quit IRC | 16:11 | |
| *** ekarlso- has quit IRC | 16:11 | |
| *** yolanda has quit IRC | 16:11 | |
| *** ndipanov has quit IRC | 16:11 | |
| *** sushils has quit IRC | 16:11 | |
| *** spzala has quit IRC | 16:11 | |
| *** monst has quit IRC | 16:11 | |
| *** pfreund has quit IRC | 16:11 | |
| *** mrodden1 has quit IRC | 16:11 | |
| *** n0ano has quit IRC | 16:11 | |
| *** mdenny has quit IRC | 16:11 | |
| *** fc__ has quit IRC | 16:11 | |
| *** krtaylor has quit IRC | 16:11 | |
| *** sushil_ is now known as sushils | 16:11 | |
| markmc | mordred, or unless you're installing fresh | 16:11 |
| smoser | vishy, no. was it here? | 16:11 |
| mordred | nope | 16:11 |
| *** monst has joined #openstack-dev | 16:11 | |
| *** spzala_ is now known as spzala | 16:11 | |
| vishy | smoser: yeah, no help adding in the log you requested | 16:11 |
| mordred | markmc: pip will literally never install the pre-release versoin string versions | 16:11 |
| mordred | markmc: unless you ask for them explicitly | 16:11 |
| markmc | mordred, ah, ok - interesting | 16:11 |
| dhellmann | jd__: the conflict we're having is in devstack installations, so maybe you're right and we should just require the newer version of kombu and then update the version in devstack | 16:12 |
| *** sushils has quit IRC | 16:12 | |
| mordred | markmc: so "pip install oslo.config" would give you 1.1 but "pip install oslo.config==1.2a1" would give you 1.2.a1 | 16:12 |
| *** neelashah1 has joined #openstack-dev | 16:12 | |
| markmc | mordred, still not exactly what we want, but better I guess | 16:12 |
| vishy | smoser: came out like this http://paste.openstack.org/show/38341/ | 16:12 |
| mordred | markmc: it's not? | 16:12 |
| markmc | mordred, we'd like to be testing e.g. trunk nova with oslo.config 1.2a1 | 16:12 |
| markmc | mordred, even if nova doesn't actually require 1.2 APIs | 16:12 |
| *** neelashah has quit IRC | 16:12 | |
| *** michchap has joined #openstack-dev | 16:13 | |
| markmc | mordred, i.e. >=1.1.0 is still an accurate reflection of what nova requires | 16:13 |
| jd__ | dhellmann: I like when you say I'm right; would you like me to push the change or will you amend? | 16:13 |
| *** alop has joined #openstack-dev | 16:13 | |
| smoser | vishy, hm.. i have to play with it. but just looking at busybox, i thought i would have had the necessary udhcp stuff. | 16:13 |
| dhellmann | jd__: you push it; I'll abandon mine | 16:13 |
| smoser | but clearly not. | 16:13 |
| mordred | markmc: I think we can handle testing trunk nova with trunk oslo.config out of band of that | 16:13 |
| mordred | markmc: but we just as importantly want to test trunk nova with latest released oslo.config, nee? | 16:14 |
| vishy | smoser: i think you need https://dev.openwrt.org/browser/trunk/package/busybox/patches/244-udhcpc_cidrroute.patch?rev=16433 or | 16:14 |
| markmc | mordred, what happens if I do oslo.config>=1.2a1 ? | 16:14 |
| vishy | smoser: https://lists.openwrt.org/pipermail/openwrt-devel/2011-January/009470.html | 16:14 |
| markmc | mordred, that works installs 1.2a1 too ? | 16:14 |
| mordred | I don't know. dstufft ^^ ? | 16:14 |
| markmc | mordred, thinking of if 1.2a2 is installed, we don't want to back down | 16:14 |
| dhellmann | markmc: yeah, that should install 1.2a1 and then when 1.2 is released switch to that | 16:14 |
| markmc | dhellmann, ok | 16:14 |
| dstufft | yes | 16:14 |
| dhellmann | I think | 16:14 |
| dstufft | the pre reelase thing works with a simple flag basically | 16:15 |
| mordred | will it install 1.2a2 if one of those comes out? | 16:15 |
| dhellmann | but if you said >=1.2 you would not get 1.2a1 | 16:15 |
| dstufft | "should pre-releases be considered" | 16:15 |
| mordred | PERFECT | 16:15 |
| *** monst has quit IRC | 16:15 | |
| dstufft | and that defaults to false, and toggles true with either a --pre flag, or including a pre rellease version in a version spec | 16:15 |
| *** yidclare has quit IRC | 16:16 | |
| dstufft | so >=1.2a1 would allow any pre-release >=1.2.a1 (or regular release) | 16:16 |
| dstufft | (For that particular package) | 16:16 |
| *** abhishekkr has joined #openstack-dev | 16:16 | |
| *** mrodden has joined #openstack-dev | 16:16 | |
| dstufft | version spec toggle is on a per package level, --pre cli flag is global | 16:16 |
| markmc | dstufft, is there a pip --consider-pre-releases command line flag ? | 16:16 |
| *** stevemar has quit IRC | 16:16 | |
| dstufft | markmc: --pre | 16:16 |
| *** ndipanov has joined #openstack-dev | 16:16 | |
| *** pixelbeat has quit IRC | 16:16 | |
| markmc | dstufft, ah, that helps too | 16:16 |
| dstufft | will tell pip to globally consider pre releases | 16:17 |
| dstufft | for all packages you're attempting to install | 16:17 |
| *** stevemar has joined #openstack-dev | 16:17 | |
| markmc | heh, globally | 16:17 |
| markmc | that includes sqlalchemy and eventlet? | 16:17 |
| dstufft | well global to that invocation | 16:17 |
| markmc | heh | 16:17 |
| markmc | yeah, I got it | 16:17 |
| markmc | we might be wary of pre-release versions of non-openstack thingies | 16:17 |
| markmc | it's definitely useful stuff, though | 16:17 |
| dhellmann | markmc: do you want to gate oslo projects on the tests in the other projects? | 16:18 |
| *** jclift has joined #openstack-dev | 16:18 | |
| *** pfreund has joined #openstack-dev | 16:18 | |
| dstufft | yea well the idea is that installing a pre release is probably not what someone who is doing `pip install foo` wants | 16:18 |
| *** sushils has joined #openstack-dev | 16:19 | |
| markmc | dhellmann, as in, don't merge an oslo.config/pbr/hacking change without checking that it doesn't break devstack and unit tests of all the other projects? | 16:19 |
| smoser | vishy, but i thought i had that. | 16:19 |
| markmc | dhellmann, yes please :) | 16:19 |
| openstackgerrit | A change was merged to openstack/python-keystoneclient: Allow keystoneclient to work with older keystone installs https://review.openstack.org/25306 | 16:19 |
| dhellmann | markmc: (git clone nova && git clone oslo.config && cd nova && tox -e py27 --notest && cd ../oslo.config && ../nova/.tox/py27/bin/python setup.py install && cd ../nova && tox -e py27) | 16:20 |
| mordred | dstufft: not that you need to consume more of our crazy, but you might find lurking in #openstack-infra interesting from time to time. that's where we wind up talking about things like pypi mirrors and crazy requirements | 16:20 |
| dhellmann | markmc: we're doing some similar stuff internally | 16:20 |
| vishy | smoser: not sure if it is route or cidrroute tbh :( | 16:20 |
| markmc | dhellmann, that would be great | 16:20 |
| vishy | iannoying that there are two of them | 16:20 |
| smoser | yeah, that was confusing. | 16:20 |
| dstufft | mordred: joined | 16:20 |
| mordred | markmc, dhellmann: we'd also talked about adding oslo.config to the devstack gate like we did for pbr, no? | 16:21 |
| *** abhishekkr has quit IRC | 16:21 | |
| mordred | (not that putting pbr in the devstack gate caught everything...) | 16:21 |
| mordred | :) | 16:21 |
| markmc | mordred, we had, but haven't done it yet | 16:21 |
| dhellmann | mordred, markmc : that would probably be simpler :-) | 16:21 |
| dhellmann | jd__: I thought you said we had to go to kombo >=2.5.10? | 16:22 |
| markmc | dhellmann, like with pbr, it won't catch issues which just affect unit tests | 16:22 |
| *** otherwiseguy has quit IRC | 16:22 | |
| mordred | dhellmann: yeah - given that we have multi-project combined testing stuff already going - we should make ample use of it | 16:22 |
| markmc | dhellmann, like the MultiConfigParser breakage with quantum | 16:22 |
| mordred | well... | 16:22 |
| markmc | dhellmann, but it would be a great start | 16:22 |
| mordred | so - the devstack-gate nodes get the starting point of all of the repos in all of the states they need | 16:22 |
| jd__ | dhellmann: no I didn't say that, I've checked against the git history of kombu, check the backlog? | 16:22 |
| mordred | you can totally write additional jobs which do not run devstack | 16:23 |
| dhellmann | jd__: sorry, I misunderstood | 16:23 |
| *** bdpayne has joined #openstack-dev | 16:23 | |
| mordred | but which do tests on the corpus of repos | 16:23 |
| jd__ | dhellmann: np :) | 16:23 |
| *** michchap has quit IRC | 16:23 | |
| mordred | so you could write a jenkins job that runs on a d-g node and which goes and does sort of what dhellman was doing above, except with all of the unittests in all of the projects | 16:23 |
| *** terry7 has joined #openstack-dev | 16:23 | |
| dhellmann | jd__: don't we have a bug for this that could be referenced in the commit message? | 16:23 |
| mordred | we've been considering a similar test for hacking - make sure we don't hose projects with new hacking releases | 16:24 |
| jd__ | dhellmann: hmm it seems we have #1134575 | 16:24 |
| jd__ | I'll ad it | 16:24 |
| *** krtaylor_ has quit IRC | 16:25 | |
| *** kbrierly has joined #openstack-dev | 16:26 | |
| dhellmann | mordred: I think markmc is saying we would want to run the unit tests for other projects as part of the devstack gates -- I guess we don't do that now? | 16:26 |
| *** xga has quit IRC | 16:26 | |
| *** savid has quit IRC | 16:26 | |
| *** savid has joined #openstack-dev | 16:27 | |
| *** souvik1 has quit IRC | 16:27 | |
| *** blues-man has quit IRC | 16:28 | |
| Alex_Gaynor | mordred: I see you removed the distribute dependency in pbr, that seems to have improved the situation somewhat: http://bpaste.net/show/106218/ keystone-client still doesn't install, but at least my venv isn't broken :) Fix is still --distribute? | 16:29 |
| dstufft | mordred: Alex_Gaynor fwiw https://gist.github.com/dstufft/5758377 | 16:31 |
| *** zoresvit has quit IRC | 16:31 | |
| mordred | markmc, jeblair: ok. the toci test shows that latest trunk commit fixes them. we still may not have fixed python-quantumclient all the way - but I'm inclined to consider that continuing work and that we should release 0.5.16 to get us out of rampant trunk breakage | 16:31 |
| *** ayoung has quit IRC | 16:32 | |
| mordred | yeah - that's still grabbing the currently-released-pbr | 16:32 |
| savid | Does anyone know how I might be able to set up a devstack that has a V2 keystone backend? | 16:32 |
| *** devoid has joined #openstack-dev | 16:33 | |
| dstufft | mordred: ok, just thought i'd try it out :) gotta go now though | 16:33 |
| mordred | dstufft: cool. thanks for you help | 16:33 |
| *** abhishekkr has joined #openstack-dev | 16:33 | |
| *** markmc has quit IRC | 16:34 | |
| Alex_Gaynor | mordred: hmm, it's producing a different result for me than yesterday, so I assumed it was with your commit? | 16:34 |
| *** pnavarro has quit IRC | 16:34 | |
| *** pnavarro has joined #openstack-dev | 16:34 | |
| mordred | Alex_Gaynor: we've released since yeseterday | 16:34 |
| *** JordanP has quit IRC | 16:34 | |
| Alex_Gaynor | mordred: ah, but not with your latest commit? | 16:34 |
| mordred | Alex_Gaynor: and I created a whole new exciting realm of breakage | 16:34 |
| Alex_Gaynor | mordred: New and exciting things! | 16:34 |
| mordred | Alex_Gaynor: no. because the latest release broke _everything_ we're extra testing this one just to make sure | 16:35 |
| * dhellmann thinks mordred never does things by halves | 16:35 | |
| Alex_Gaynor | dhellmann: fancy seeing you here | 16:35 |
| mordred | Alex_Gaynor: also, dstufft has pointed me in some good directions of simplifying the problem space | 16:35 |
| *** kaushikc has joined #openstack-dev | 16:35 | |
| dhellmann | Alex_Gaynor: I had the same thought about you. :-) | 16:35 |
| Alex_Gaynor | mordred: that's great, is there anything I can do to help with getting the fixed pbr out the door | 16:35 |
| *** schwicht has quit IRC | 16:36 | |
| *** xqueralt is now known as xqueralt-afk | 16:36 | |
| *** krtaylor has joined #openstack-dev | 16:38 | |
| *** abhishekkr has quit IRC | 16:38 | |
| *** drewlander has joined #openstack-dev | 16:39 | |
| *** neelashah has joined #openstack-dev | 16:39 | |
| *** lucasagomes has quit IRC | 16:40 | |
| *** pmyers has quit IRC | 16:40 | |
| *** neelashah has quit IRC | 16:40 | |
| *** neelashah1 has quit IRC | 16:40 | |
| *** portante has joined #openstack-dev | 16:40 | |
| *** kevinconway has quit IRC | 16:41 | |
| *** neelashah has joined #openstack-dev | 16:41 | |
| mordred | dhellmann: does my statement to markmc and jeblair above make sense to you? | 16:41 |
| *** pmyers has joined #openstack-dev | 16:41 | |
| *** monst has joined #openstack-dev | 16:42 | |
| *** sthaha has quit IRC | 16:42 | |
| mordred | dhellmann: (looking for another voice double check my thinking here) | 16:42 |
| *** eglynn has quit IRC | 16:42 | |
| dhellmann | mordred: the one about releasing 0.5.16? | 16:43 |
| mordred | dhellmann: yes | 16:43 |
| *** pnavarro has quit IRC | 16:44 | |
| *** jtomasek has joined #openstack-dev | 16:44 | |
| dhellmann | mordred: I agree that fixing something makes the situation better. I'm not sure what "toci tests" are | 16:44 |
| mordred | dhellmann: oh, derekh has some CI tests he runs on the tripleo stuff | 16:44 |
| mordred | dhellmann: one of the things it happens to do is install pbr from git | 16:44 |
| *** pnavarro has joined #openstack-dev | 16:44 | |
| mordred | dhellmann: so it caught the error we're seeing originally | 16:45 |
| dhellmann | mordred: but if 0.5.16 helps most of the projects, (ah) then I say we go for it. There are a lot of numbers after 16 than we can use for update releases. | 16:45 |
| *** AnilV4 has quit IRC | 16:46 | |
| *** pfreund has quit IRC | 16:46 | |
| mordred | dhellmann: ok. actually, I tested python-quantumclient too and I actually think we're fixed there as well | 16:48 |
| dhellmann | mordred: even better | 16:48 |
| mordred | dhellmann: but I've learned a new thing | 16:48 |
| *** monst has quit IRC | 16:48 | |
| dhellmann | logoff after committing a breaking change? :-) | 16:48 |
| *** monst has joined #openstack-dev | 16:48 | |
| mordred | dhellmann: (git clone nova && git clone oslo.config && cd nova && tox -e py27 --notest && cd ../oslo.config && ../nova/.tox/py27/bin/python setup.py install && cd ../nova && tox -e py27) has a potential flaw | 16:48 |
| dhellmann | oh? | 16:49 |
| mordred | dhellmann: (git clone nova && git clone oslo.config && cd nova && tox -e py27 --notest && ../nova/.tox/py27/bin/pip install git+file///../oslo.config && cd ../nova && tox -e py27) has a potential flaw | 16:49 |
| mordred | dhellmann: cd-ing into the oslo.config dir can cause the setup_requires to download an egg and therefore not actually get a depend instlaled into the venv | 16:49 |
| *** michchap has joined #openstack-dev | 16:50 | |
| mordred | because of easy_install suck | 16:50 |
| *** rharwood has quit IRC | 16:50 | |
| dhellmann | ah | 16:50 |
| *** salv-orlando has quit IRC | 16:50 | |
| *** abhishekkr has joined #openstack-dev | 16:50 | |
| *** BobBall has quit IRC | 16:52 | |
| *** katylava is now known as katylavaway | 16:52 | |
| dhellmann | mordred: if we're doing the tests as part of the oslo gate, it would probably be simpler to install the version of the code checked out by jenkins (whatever is in the working directory) instead of cloning a fresh copy | 16:52 |
| *** networkstatic has quit IRC | 16:52 | |
| mordred | dhellmann: definitely | 16:53 |
| *** harlowja has joined #openstack-dev | 16:53 | |
| *** vipul|away is now known as vipul | 16:53 | |
| dhellmann | git clone nova {toxenvdir}/nova && pip install -r {toxenvdir}/nova/requirements.txt -r {toxenvdir}/nova/test-requirements.txt && (cd {toxenvdir}/nova && pip install -e . && python testr) | 16:55 |
| *** seanrob has quit IRC | 16:56 | |
| *** seanrob has joined #openstack-dev | 16:56 | |
| *** seanrob_ has joined #openstack-dev | 16:57 | |
| *** vipul is now known as vipul|away | 16:57 | |
| *** monst has quit IRC | 16:57 | |
| *** abhishekkr has quit IRC | 16:57 | |
| *** johnthetubaguy1 has joined #openstack-dev | 16:59 | |
| *** marun has joined #openstack-dev | 16:59 | |
| *** derekh has quit IRC | 16:59 | |
| *** abhisri has joined #openstack-dev | 17:00 | |
| *** nati_ueno has joined #openstack-dev | 17:00 | |
| *** yidclare has joined #openstack-dev | 17:00 | |
| *** johnthetubaguy has quit IRC | 17:01 | |
| *** seanrob has quit IRC | 17:01 | |
| *** michchap has quit IRC | 17:01 | |
| *** johnthetubaguy1 has quit IRC | 17:02 | |
| *** pixelbeat has joined #openstack-dev | 17:02 | |
| *** portante has quit IRC | 17:03 | |
| *** rushiagr has joined #openstack-dev | 17:04 | |
| *** tiamar has joined #openstack-dev | 17:05 | |
| *** devvesa has quit IRC | 17:06 | |
| *** shang has quit IRC | 17:06 | |
| *** changbl has joined #openstack-dev | 17:07 | |
| savid | Anyone familiar with devstack? I'm getting 404 errors when calling keystone v2 API calls | 17:08 |
| *** SergeyLukjanov has quit IRC | 17:10 | |
| kaushikc | savid: is your stack running? | 17:10 |
| *** abhishekkr has joined #openstack-dev | 17:10 | |
| *** eglynn has joined #openstack-dev | 17:10 | |
| kaushikc | if so join the screen session and see what went wrong | 17:10 |
| *** wu_wenxiang has joined #openstack-dev | 17:11 | |
| *** otherwiseguy has joined #openstack-dev | 17:11 | |
| *** gyee has joined #openstack-dev | 17:12 | |
| *** sleepsonthefloor has joined #openstack-dev | 17:12 | |
| savid | kaushikc, yeah, here's the output from the keystone screen: http://dpaste.com/1229493/ You'll see that the last call is a 404. | 17:12 |
| *** mrunge has joined #openstack-dev | 17:13 | |
| *** dolphm has joined #openstack-dev | 17:13 | |
| *** abhishekkr has quit IRC | 17:14 | |
| savid | Oh, I just realized.. something's calling /v2/users instead of /v2.0/users. That's probably it. | 17:16 |
| savid | Hmm, ok I just fixed it so that it's calling /v2.0/users, however I'm still getting a 404. http://192.168.33.10:5000/v2.0/users | 17:19 |
| *** Tross has joined #openstack-dev | 17:19 | |
| *** sandywalsh has quit IRC | 17:19 | |
| *** terryh has quit IRC | 17:19 | |
| *** jcoufal has quit IRC | 17:20 | |
| *** yjiang5_ has joined #openstack-dev | 17:20 | |
| *** esp has left #openstack-dev | 17:20 | |
| *** yjiang5 has left #openstack-dev | 17:20 | |
| mordred | jd__: will you test the things that you test to see if the new pbr fixes your world? | 17:20 |
| *** eglynn has quit IRC | 17:21 | |
| *** spzala has quit IRC | 17:22 | |
| *** jtomasek has quit IRC | 17:23 | |
| *** morgank has joined #openstack-dev | 17:23 | |
| mordred | dhellmann: ^^ | 17:24 |
| mordred | jd__, dhellmann: fix is in the wild and synced to our mirrors. local testing shows that python-quantumclient is fixed now too - although that's being further tested | 17:24 |
| mordred | dhellmann: do you have control over cliff? | 17:24 |
| *** rharwood has joined #openstack-dev | 17:25 | |
| jd__ | mordred: rechecking | 17:26 |
| *** spzala has joined #openstack-dev | 17:26 | |
| *** zodiak has quit IRC | 17:27 | |
| *** abhishekkr has joined #openstack-dev | 17:27 | |
| *** michchap has joined #openstack-dev | 17:27 | |
| Alex_Gaynor | mordred: can confirm that pip install python-keystoneclient works, yay! A++ would spend far too long investigating again ;) | 17:28 |
| mordred | Alex_Gaynor: WOOT! | 17:28 |
| *** HenryG has quit IRC | 17:28 | |
| mordred | jeez. what a fun time this has been | 17:28 |
| Alex_Gaynor | mordred: anyone remember what I was trying to do at the beginning? | 17:29 |
| *** cp16net is now known as cp16net|away | 17:29 | |
| mordred | Alex_Gaynor: nope. wait - you weren't trying to become a pbr core hacker? | 17:29 |
| Alex_Gaynor | that doesn't sound right | 17:29 |
| *** cmark has joined #openstack-dev | 17:30 | |
| * mordred hypnotizes Alex_Gaynor | 17:30 | |
| *** amerine has quit IRC | 17:31 | |
| *** timello_ has quit IRC | 17:31 | |
| *** sandywalsh has joined #openstack-dev | 17:32 | |
| openstackstatus | NOTICE: pbr 0.5.16 has been released and the gate should be back in business | 17:32 |
| *** ChanServ changes topic to "OpenStack development || Support is in #openstack" | 17:32 | |
| *** jongleur has joined #openstack-dev | 17:32 | |
| *** JonnyNomad has quit IRC | 17:33 | |
| savid | kaushikc, is there another screen that might tell me more info? The keystone service just shows the HTTP response being 404, which is nothing helpful. | 17:33 |
| chuckieb | Our group is struggling with adding a new hypervisor for the zSeries mainframe to OpenStack... too much code for review vs not enough function | 17:33 |
| *** rcrit has quit IRC | 17:33 | |
| *** amerine has joined #openstack-dev | 17:33 | |
| chuckieb | Would you recommend going with the function to support spawn, even if it is a large first submit? | 17:34 |
| *** rushiagr has quit IRC | 17:34 | |
| *** rkukura has quit IRC | 17:34 | |
| *** Mandell has joined #openstack-dev | 17:36 | |
| *** rushiagr has joined #openstack-dev | 17:36 | |
| *** enikanorov has joined #openstack-dev | 17:36 | |
| mordred | dhellmann: https://github.com/dreamhost/cliff/pull/35 | 17:37 |
| mordred | chuckieb: interesting! so, I know that nova has been relucatant recently to take new hypervisors without a story on how we're going to test it | 17:37 |
| *** souvik has joined #openstack-dev | 17:38 | |
| *** enikanorov has quit IRC | 17:38 | |
| chuckieb | True. Not everyone has a z Series hardware around | 17:38 |
| mordred | chuckieb: perhaps we should have a chat about what that would look like? (then even a large first commit might not be _as_ terrible) | 17:38 |
| mordred | chuckieb: well, and specifically, the openstack CI infrastructure most certainly does not :) | 17:38 |
| *** enikanorov has joined #openstack-dev | 17:38 | |
| *** michchap has quit IRC | 17:39 | |
| *** pabelanger has quit IRC | 17:39 | |
| chuckieb | To get OpenStack support we used the Rest-API interface in the xCAT product that run on z series (and other platforms) | 17:39 |
| *** abhishekkr has quit IRC | 17:39 | |
| *** timello has joined #openstack-dev | 17:39 | |
| chuckieb | the zSeries zVM hypervisor does not have a libvert interface, so we decided on interfacing to the xCAT code | 17:40 |
| cmark | Any keystone people around? Am I to expect that some of the tests executed by run_tests.sh are going to fail? | 17:41 |
| *** Ryan_Lane has joined #openstack-dev | 17:41 | |
| dolphm | cmark: https://bugs.launchpad.net/keystone/+bug/1172794 | 17:41 |
| uvirtbot | Launchpad bug 1172794 in keystone "Some nose.exc.SkipTest() rendered as ERROR's" [Low,Confirmed] | 17:41 |
| chuckieb | The xCAT code is in sourceforge. This makes it interesting coding in 3 separate places, OpenStack, Sourceforge, and internal zVM | 17:41 |
| *** wu_wenxiang has quit IRC | 17:42 | |
| *** pmathews has quit IRC | 17:42 | |
| cmark | dolphm: thanks. I'll check it out | 17:42 |
| *** pmathews has joined #openstack-dev | 17:42 | |
| *** seanrob_ has quit IRC | 17:42 | |
| dolphm | cmark: fix appreciated :) i haven't looked too much into it | 17:42 |
| *** jbresnah has joined #openstack-dev | 17:44 | |
| *** eglynn has joined #openstack-dev | 17:44 | |
| *** datsun180b has quit IRC | 17:45 | |
| *** sandywalsh has quit IRC | 17:45 | |
| mordred | cmark: hehe | 17:45 |
| mordred | gah | 17:45 |
| mordred | chuckieb: hehe | 17:45 |
| cmark | heh | 17:45 |
| mordred | chuckieb: so - it seems like there might be two different conversations we might want to have | 17:46 |
| mordred | chuckieb: because we've got someone from IBM from the xCAT team working on Ironic, which is the baremetal provisioning code stuff for openstack | 17:46 |
| mordred | chuckieb: so it's possible that there might be multiple different ways to interface with this problem | 17:47 |
| *** yolanda_ has quit IRC | 17:47 | |
| *** READ10 has quit IRC | 17:47 | |
| mordred | devananda: ^^ | 17:47 |
| mordred | devananda: when you awaken or whatnot, might be an interesting conversation to draw you in to | 17:48 |
| *** schwicht has joined #openstack-dev | 17:49 | |
| kaushikc | russellb: around? | 17:49 |
| cmark | dolphm: I don't think this is the same issue as described in those bugs. I'm getting actual failures. FWIW, I'm using the deb-source package from ubuntu cloud archive. Should I be able to download the package and run run_tests.sh on a clean system and expect it to work? | 17:50 |
| dhellmann | mordred: on it (sorry, was at lunch) | 17:51 |
| dolphm | cmark: yes... can you paste what you're seeing? | 17:51 |
| openstackgerrit | A change was merged to openstack/python-swiftclient: Add option to disable SSL compression https://review.openstack.org/20029 | 17:51 |
| *** mkollaro has quit IRC | 17:51 | |
| *** abhishekkr has joined #openstack-dev | 17:52 | |
| cmark | dolphm: http://paste.ubuntu.com/5755660/ | 17:52 |
| cmark | the database file /var/lib/keystone/keystone.db appears to get created. But there are no tables. | 17:53 |
| dolphm | cmark: that's not the db that should be used by tests though | 17:53 |
| dolphm | cmark: that's created by keystone-all and managed by db_sync | 17:54 |
| *** cp16net|away is now known as cp16net | 17:54 | |
| *** maoy has joined #openstack-dev | 17:55 | |
| *** terryh has joined #openstack-dev | 17:55 | |
| cmark | if i run run_tests.sh without -x and let it get past this test, eventually, some later test populates the db. If I run run_tests.sh again, it appears to pass this test. | 17:55 |
| *** eglynn has quit IRC | 17:55 | |
| *** abhishekkr has quit IRC | 17:56 | |
| *** nati_uen_ has joined #openstack-dev | 17:57 | |
| *** nati_ueno has quit IRC | 17:57 | |
| *** sandywalsh has joined #openstack-dev | 17:58 | |
| *** eglynn has joined #openstack-dev | 17:58 | |
| *** abhisri has quit IRC | 17:58 | |
| *** jruzicka has quit IRC | 17:58 | |
| *** egallen has quit IRC | 17:59 | |
| *** ifarkas has quit IRC | 17:59 | |
| *** henrynash has joined #openstack-dev | 17:59 | |
| *** nachi__ has joined #openstack-dev | 17:59 | |
| *** abhisri has joined #openstack-dev | 17:59 | |
| *** pmcg_ is now known as pmcg | 18:00 | |
| *** pmcg has quit IRC | 18:00 | |
| *** pmcg has joined #openstack-dev | 18:01 | |
| *** benonsoftware has quit IRC | 18:01 | |
| *** Mandell has quit IRC | 18:01 | |
| *** eglynn has quit IRC | 18:02 | |
| *** SergeyLukjanov has joined #openstack-dev | 18:02 | |
| *** tkammer has joined #openstack-dev | 18:02 | |
| *** RenssegNevets has joined #openstack-dev | 18:02 | |
| *** marun has quit IRC | 18:03 | |
| *** michchap has joined #openstack-dev | 18:06 | |
| *** krtaylor has quit IRC | 18:06 | |
| dhellmann | mordred: https://pypi.python.org/pypi/cliff/1.4 | 18:08 |
| dhellmann | mordred: https://review.openstack.org/#/c/32598/1 | 18:08 |
| mordred | dhellmann: awesome. thanks! +2'd | 18:08 |
| *** Mandell has joined #openstack-dev | 18:09 | |
| chmouel | belliott: ping | 18:11 |
| *** jbresnah has quit IRC | 18:12 | |
| *** atiwari has joined #openstack-dev | 18:13 | |
| openstackgerrit | A change was merged to openstack/ceilometer: Fix requirements https://review.openstack.org/32577 | 18:16 |
| *** michchap has quit IRC | 18:17 | |
| *** mmagr has joined #openstack-dev | 18:17 | |
| *** Tross has quit IRC | 18:17 | |
| openstackgerrit | A change was merged to openstack/cinder: Integrate oslo's periodic tasks. https://review.openstack.org/32402 | 18:17 |
| openstackgerrit | A change was merged to openstack/nova: Nova instance group DB support https://review.openstack.org/28880 | 18:17 |
| *** tkammer has quit IRC | 18:18 | |
| *** jbresnah has joined #openstack-dev | 18:19 | |
| *** comay has joined #openstack-dev | 18:19 | |
| *** vartom19 has joined #openstack-dev | 18:20 | |
| openstackgerrit | A change was merged to openstack/oslo-incubator: Raise a DeprecationWarning when openstack.common.deprecated is imported https://review.openstack.org/32561 | 18:20 |
| *** melwitt has joined #openstack-dev | 18:21 | |
| *** psedlak has quit IRC | 18:22 | |
| cmark | dolphm: what db are the tests supposed to use then? | 18:23 |
| dolphm | cmark: it's done in memory as much as possible, but check tests/backend_sql_disk.conf for the actual config and where it's used | 18:24 |
| *** jbresnah has quit IRC | 18:25 | |
| openstackgerrit | A change was merged to openstack/ceilometer: Use console scripts for ceilometer-api https://review.openstack.org/31747 | 18:27 |
| openstackgerrit | A change was merged to openstack/ceilometer: service: run common initialization stuff https://review.openstack.org/31046 | 18:27 |
| openstackgerrit | A change was merged to openstack/nova: xenapi: ensure instance metadata always injected into xenstore https://review.openstack.org/32132 | 18:28 |
| openstackgerrit | A change was merged to openstack/ceilometer: Remove useless imports, reenable F401 checks https://review.openstack.org/31051 | 18:28 |
| *** salv-orlando has joined #openstack-dev | 18:28 | |
| *** krtaylor has joined #openstack-dev | 18:28 | |
| openstackgerrit | A change was merged to openstack/ceilometer: fix session connection https://review.openstack.org/31664 | 18:28 |
| *** RenssegNevets has quit IRC | 18:29 | |
| *** alop has quit IRC | 18:30 | |
| *** kagan has joined #openstack-dev | 18:30 | |
| *** noslzzp has quit IRC | 18:30 | |
| *** kevinconway has joined #openstack-dev | 18:30 | |
| *** alop has joined #openstack-dev | 18:31 | |
| *** datsun180b has joined #openstack-dev | 18:31 | |
| *** kaushikc has quit IRC | 18:31 | |
| *** eglynn has joined #openstack-dev | 18:31 | |
| *** yolanda_ has joined #openstack-dev | 18:31 | |
| openstackgerrit | A change was merged to openstack/keystone: Move coverage output dir for Jenkins. https://review.openstack.org/31913 | 18:32 |
| *** shardy is now known as shardy_afk | 18:32 | |
| openstackgerrit | A change was merged to openstack/ceilometer: Corrected path for test requirements in docs https://review.openstack.org/32434 | 18:33 |
| openstackgerrit | A change was merged to openstack/ceilometer: Fix some typo in documentation https://review.openstack.org/32111 | 18:33 |
| *** ladquin_afk is now known as ladquin | 18:34 | |
| sc68cal | Hey - is there a fix for the PBR stuff? I've got a broken devstack | 18:34 |
| *** varora200 has joined #openstack-dev | 18:35 | |
| *** sushils has quit IRC | 18:35 | |
| clarkb | sc68cal: yes it was recently merged and pushed to pypi | 18:35 |
| *** derekh has joined #openstack-dev | 18:36 | |
| sc68cal | clarkb: thx - let me destroy my vagrant vm and rebuild | 18:38 |
| *** dims has quit IRC | 18:38 | |
| *** terryh has quit IRC | 18:39 | |
| *** rushiagr has quit IRC | 18:40 | |
| *** rushiagr has joined #openstack-dev | 18:41 | |
| *** kaushikc has joined #openstack-dev | 18:42 | |
| openstackgerrit | A change was merged to openstack/oslo-incubator: Enable hacking H404 test. https://review.openstack.org/31511 | 18:43 |
| *** mkollaro has joined #openstack-dev | 18:43 | |
| *** michchap has joined #openstack-dev | 18:44 | |
| *** novas0x2a|laptop has joined #openstack-dev | 18:44 | |
| *** dims has joined #openstack-dev | 18:45 | |
| *** kaushikc has quit IRC | 18:45 | |
| openstackgerrit | A change was merged to openstack/keystone: Replace openstack-common with oslo in docs https://review.openstack.org/32375 | 18:46 |
| *** rushiagr has quit IRC | 18:46 | |
| *** electrichead has joined #openstack-dev | 18:49 | |
| *** michchap has quit IRC | 18:50 | |
| *** derekh has quit IRC | 18:50 | |
| *** colinmcnamara1 has joined #openstack-dev | 18:50 | |
| *** abhisri has quit IRC | 18:51 | |
| *** networkstatic has joined #openstack-dev | 18:52 | |
| *** colinmcnamara has quit IRC | 18:52 | |
| sc68cal | clarkb: yep - thanks | 18:53 |
| clarkb | np | 18:53 |
| *** _ffio_ has quit IRC | 18:54 | |
| *** mlavalle has joined #openstack-dev | 18:54 | |
| *** chuckieb has quit IRC | 18:57 | |
| *** katylavaway is now known as katylava | 18:57 | |
| *** jongleur has quit IRC | 18:58 | |
| *** derekh has joined #openstack-dev | 18:59 | |
| *** mkollaro has quit IRC | 18:59 | |
| dstufft | mordred: +1 on simplfying :D | 18:59 |
| mordred | dstufft: ++ | 18:59 |
| *** atiwari has quit IRC | 18:59 | |
| mordred | dstufft: I've just pushed changes up to all of the openstack projects removing direct depends on distribute | 19:00 |
| gyee | dolphm, henrynash, difference between enum and multiple boolean is that with enum, you can choice one | 19:00 |
| dstufft | mordred: awesome :) | 19:00 |
| dolphm | gyee: just like you can choose a state by setting booleans | 19:00 |
| dolphm | gyee: what was your full enumeration of states? i'd like to map the two solutions togetehr | 19:00 |
| henrynash | gyee: { "inherited to": ["domains", "projects"] } can have both domains and projects defined…so it's not an enum | 19:01 |
| gyee | dolphm, { "inherited to": ["domains", "projects"] } looks a lot cleaner | 19:01 |
| dolphm | gyee: that's not an enumeration | 19:01 |
| dolphm | gyee: how does 'global' fit there? | 19:01 |
| *** pabelanger has joined #openstack-dev | 19:02 | |
| *** vipul|away is now known as vipul | 19:02 | |
| henrynash | gyee: its a collection of booleans (i.e. defined or not) | 19:02 |
| sc68cal | Hey - nova commit 584b2fb4b0a1968699f8d2dd4b0a2af67e7dca8c breaks devstack | 19:02 |
| gyee | you can add "global" in the list to choose from | 19:02 |
| sc68cal | no more bin dir to start up nova-api | 19:02 |
| henrynash | gyee: no you can't!! "inherited to" [global]…what does that mean? | 19:03 |
| gyee | henrynash, means you got the role regardless of token scope | 19:03 |
| openstackgerrit | A change was merged to openstack/ceilometer: Add instance_scheduled in entry points https://review.openstack.org/32108 | 19:03 |
| dolphm | henrynash: +1 | 19:04 |
| henrynash | gyee: woah…..wrong place to put that….a role_def in and of itself cannot grant/assign a role…we must still have you assign it | 19:05 |
| gyee | henrynash, then what's the point of have that attribute in the role def? | 19:05 |
| gyee | having | 19:05 |
| *** eglynn has quit IRC | 19:06 | |
| dolphm | henrynash: gyee: can we first agree that these are the use cases we're addressing, using today's terminology? https://etherpad.openstack.org/keystone-role-inheritance | 19:06 |
| bknudson | henrynash: with the new design, if I change the role definition it doesn't change any existing assignments ? | 19:06 |
| henrynash | bknudson: correct, they are independent, by design | 19:06 |
| bknudson | I thought before the assignments were automatic. | 19:06 |
| henrynash | bknudson: the roledef just defines HOW a role assignment should be interpreted | 19:07 |
| gyee | henrynash, then lets not change role def, and concentrate on role assignment instead | 19:07 |
| bknudson | so if I assign a role to a user on a project, and that role is inherits, nothing different there | 19:07 |
| henrynash | gyee: AAAAGGHHHH….that's where I started……and got push back that using the roledef was the way to go… | 19:08 |
| bknudson | but if I assign role to a user on a domain, and it inherits, now it's also assigned to the rest. | 19:08 |
| dolphm | bknudson: if inherits_to_projects is mutable for example, and you PATCH'd that value on the role itself, i would expect keystone to start applying the new inheritance rules to new tokens | 19:08 |
| *** salv-orlando has quit IRC | 19:08 | |
| bknudson | dolphm: I was thinking that that was not the case. | 19:08 |
| henrynash | dolphm: agreed | 19:08 |
| dolphm | bknudson: i'd be happy if we said they were immutable in v3.1 to avoid the whole conversation for now :P | 19:09 |
| henrynash | dolphm: not sure why the have to be immutable…we calculate the effective roles a user has a auth anyway today (because of groups)…so there is no change with ths | 19:09 |
| *** abhisri has joined #openstack-dev | 19:10 | |
| dolphm | henrynash: my concern is really that PATCH'ing a role means suddenly revoking lots of tokens | 19:11 |
| henrynash | dolphm: it's true | 19:12 |
| sc68cal | false alarm, some really odd stuff going on in my devstack | 19:14 |
| sc68cal | nova api didn't come up - complained about mising bin/nova-api | 19:14 |
| sc68cal | ran again and it came up I think.... | 19:14 |
| *** dkuffner has joined #openstack-dev | 19:15 | |
| bknudson | henrynash: dolphm: If they can't PATCH, they'll just DELETE and POST. | 19:15 |
| *** LinuxJedi has quit IRC | 19:16 | |
| *** LinuxJedi has joined #openstack-dev | 19:16 | |
| dolphm | bknudson: yeah, but we already handle the consequences of delete_role... elegantly handling the consequences of change_semantics_of_role is complicated and i'm lazy and know that we won't get it quite right the first time | 19:16 |
| *** michchap has joined #openstack-dev | 19:17 | |
| *** ladquin is now known as ladquin_brb | 19:17 | |
| *** andrew_plunk has quit IRC | 19:18 | |
| henrynash | dolphm: hmm, but if we don't let them patch, deleting and recreating a role will (I think) delete all related assignments using that role…which you would have to re-create. Not good. | 19:18 |
| *** READ10 has joined #openstack-dev | 19:19 | |
| dolphm | i kind of wish we had stuck with composite keys for projects in the spec :( /domains/{domain_id}/projects/{project_id}/ because then we could do things like PUT /domains/{domain_id}/projects/*/users/{user_id}/roles/{role_id} | 19:19 |
| *** benonsoftware has joined #openstack-dev | 19:19 | |
| *** dkuffner has quit IRC | 19:19 | |
| dolphm | henrynash: good point | 19:20 |
| *** andrew_plunk has joined #openstack-dev | 19:20 | |
| gyee | I like the * mechanism | 19:21 |
| bknudson | * doesn't seem very rest-y ... is it a resource? can I GET it? | 19:21 |
| dolphm | bknudson: not really :( | 19:22 |
| dolphm | bknudson: maybe get back a list of Locations of new resources? | 19:22 |
| katylava | sure, it could be an index | 19:22 |
| henrynash | dolphm: which is why I had /inherited origionally | 19:22 |
| dolphm | henrynash: how do you feel between /inherited vs using role attributes at this point? | 19:23 |
| bknudson | henrynash: I like /inherited. | 19:23 |
| *** atiwari has joined #openstack-dev | 19:23 | |
| *** mrunge has quit IRC | 19:24 | |
| *** kagan has quit IRC | 19:24 | |
| gyee | yeah, role attributes seem wrong | 19:24 |
| *** marun has joined #openstack-dev | 19:25 | |
| *** eglynn has joined #openstack-dev | 19:25 | |
| henrynash | dolphm: even if we used /inherited we still have a problem if we wanted to, say, have the domain inherited roles….. | 19:25 |
| henrynash | dolphm: how to you assign to the "root" of all domains? | 19:25 |
| dolphm | henrynash: did i miss a conversation somewhere about applying domain hierarchies to all this? | 19:26 |
| *** mrunge has joined #openstack-dev | 19:26 | |
| henrynash | dolphm: no, there was no conversation on that…not implying domain hierarchy here, just how would you assign something to all domains? | 19:27 |
| *** vartom110 has joined #openstack-dev | 19:27 | |
| *** pnavarro has quit IRC | 19:28 | |
| henrynash | dolphm: Ok here's an idea (which I partially suggested before) | 19:28 |
| *** markmcclain has joined #openstack-dev | 19:28 | |
| *** nachi__ has quit IRC | 19:28 | |
| openstackgerrit | A change was merged to openstack/python-openstackclient: Remove explicit distribute depend. https://review.openstack.org/32614 | 19:29 |
| henrynash | PUT /roles/roleid/users/userid/domains/domainID/inherit_to_projects | 19:29 |
| *** michchap has quit IRC | 19:29 | |
| *** boris-42 has joined #openstack-dev | 19:29 | |
| henrynash | PUT /roles/roleid/users/userid/domains/inherit_to_projects (this would inherit to all projects in all domains) | 19:29 |
| *** vartom19 has quit IRC | 19:29 | |
| *** pabelanger has quit IRC | 19:29 | |
| henrynash | UT /roles/roleid/users/userid/domains/inherit_to_domains (this would inherit to all domains) | 19:29 |
| bknudson | henrynash: there's a potential name clash there if I give my domain ID the name "inherit_to_projects" | 19:30 |
| *** souvik1 has joined #openstack-dev | 19:30 | |
| henrynash | bknudson: hmm, damn :-) | 19:30 |
| *** ayoung has joined #openstack-dev | 19:30 | |
| henrynash | (this is kind of why we bailed on the api idea, it got all messy!) | 19:31 |
| russellb | dolphm: dude, <3 next-review | 19:31 |
| russellb | using the heck out of it | 19:31 |
| *** souvik has quit IRC | 19:31 | |
| dolphm | russellb: lol me too, anyone else? | 19:31 |
| bknudson | dolphm: next-review complained that my key was encrypted or something. | 19:31 |
| henrynash | dolphm, gyee: Ok, let me mull on it tonight | 19:32 |
| bknudson | paramiko.PasswordRequiredException: Private key file is encrypted | 19:32 |
| dolphm | bknudson: your ssh key is password protected? | 19:32 |
| bknudson | of course, I unlock it with my ssh-agent. | 19:32 |
| russellb | once ssh-agent is running, should work | 19:33 |
| russellb | guess we could catch that exception and add some handy help text about that | 19:33 |
| *** danwent has quit IRC | 19:33 | |
| *** mmagr has quit IRC | 19:33 | |
| bknudson | ssh git commands work but for some reason maybe paramiko on my system doesn't work with it. | 19:33 |
| dolphm | bknudson: i'd have to try it... i imagine next-review should be able to handle that error | 19:34 |
| *** seanrob has joined #openstack-dev | 19:34 | |
| openstackgerrit | A change was merged to openstack/cinder: Add availability zone checking in the api service https://review.openstack.org/32422 | 19:36 |
| openstackgerrit | A change was merged to openstack/swift: Fixed Bug 1187200 https://review.openstack.org/31584 | 19:36 |
| uvirtbot | Launchpad bug 1187200 in swift "Bug with mismatched expiring_objects_container_divisor settings" [Undecided,In progress] https://launchpad.net/bugs/1187200 | 19:36 |
| openstackgerrit | A change was merged to openstack/swift: Local read affinity for GET/HEAD requests. https://review.openstack.org/32076 | 19:36 |
| *** Tross has joined #openstack-dev | 19:37 | |
| *** alop has quit IRC | 19:37 | |
| *** vipul is now known as vipul|away | 19:37 | |
| *** sarob has joined #openstack-dev | 19:37 | |
| *** Tross has quit IRC | 19:37 | |
| *** alop has joined #openstack-dev | 19:38 | |
| *** READ10 has quit IRC | 19:38 | |
| *** sarob has quit IRC | 19:40 | |
| *** danwent has joined #openstack-dev | 19:40 | |
| *** timello has quit IRC | 19:41 | |
| *** timello has joined #openstack-dev | 19:42 | |
| *** kevinconway has left #openstack-dev | 19:42 | |
| *** eglynn has quit IRC | 19:44 | |
| *** mkollaro has joined #openstack-dev | 19:46 | |
| *** alexpilotti has quit IRC | 19:47 | |
| openstackgerrit | A change was merged to openstack/nova: Rename functions in nova.compute.flavors from instance_type https://review.openstack.org/29020 | 19:49 |
| *** yolanda_ has quit IRC | 19:49 | |
| *** avishay has quit IRC | 19:53 | |
| *** lorin1 has joined #openstack-dev | 19:53 | |
| *** michchap has joined #openstack-dev | 19:55 | |
| *** topol has quit IRC | 19:55 | |
| *** marun has quit IRC | 19:57 | |
| *** ayoung has quit IRC | 19:57 | |
| *** colinmcnamara1 has quit IRC | 19:58 | |
| *** Tross has joined #openstack-dev | 19:59 | |
| *** benner has quit IRC | 19:59 | |
| *** Tross has quit IRC | 20:00 | |
| *** marun has joined #openstack-dev | 20:00 | |
| *** pcm__ has quit IRC | 20:00 | |
| *** drewlander has quit IRC | 20:00 | |
| *** Ruetobas has quit IRC | 20:01 | |
| *** colinmcnamara has joined #openstack-dev | 20:01 | |
| *** marun has quit IRC | 20:01 | |
| *** Tross has joined #openstack-dev | 20:04 | |
| *** benner has joined #openstack-dev | 20:05 | |
| *** michchap has quit IRC | 20:05 | |
| *** souvik1 has quit IRC | 20:06 | |
| *** eglynn has joined #openstack-dev | 20:06 | |
| *** adalbas has quit IRC | 20:07 | |
| *** davidlenwell has quit IRC | 20:08 | |
| *** bknudson has quit IRC | 20:08 | |
| *** souvik has joined #openstack-dev | 20:09 | |
| *** adalbas has joined #openstack-dev | 20:10 | |
| *** ayoung has joined #openstack-dev | 20:11 | |
| *** vartom110 has quit IRC | 20:12 | |
| *** eglynn has quit IRC | 20:13 | |
| *** eglynn has joined #openstack-dev | 20:15 | |
| *** boris-42 has quit IRC | 20:15 | |
| *** atiwari has quit IRC | 20:15 | |
| *** dolphm has quit IRC | 20:17 | |
| *** vipul|away is now known as vipul | 20:18 | |
| *** dprince has quit IRC | 20:19 | |
| *** gyee has quit IRC | 20:20 | |
| *** marun has joined #openstack-dev | 20:25 | |
| *** colinmcnamara has quit IRC | 20:26 | |
| *** eglynn has quit IRC | 20:26 | |
| *** timello has quit IRC | 20:28 | |
| *** seanrob has quit IRC | 20:28 | |
| *** seanrob has joined #openstack-dev | 20:28 | |
| *** vartom110 has joined #openstack-dev | 20:29 | |
| *** dolphm has joined #openstack-dev | 20:31 | |
| *** michchap has joined #openstack-dev | 20:32 | |
| *** seanrob has quit IRC | 20:33 | |
| *** egallen has joined #openstack-dev | 20:34 | |
| *** timello has joined #openstack-dev | 20:34 | |
| openstackgerrit | A change was merged to openstack/python-openstackclient: Add volume backup commands https://review.openstack.org/31501 | 20:34 |
| *** danwent has quit IRC | 20:36 | |
| *** marun has quit IRC | 20:36 | |
| *** marun has joined #openstack-dev | 20:37 | |
| openstackgerrit | A change was merged to openstack/nova: Return Customer's Quota Usage through Admin API https://review.openstack.org/27468 | 20:39 |
| *** seanrob has joined #openstack-dev | 20:39 | |
| *** souvik1 has joined #openstack-dev | 20:40 | |
| *** noslzzp has joined #openstack-dev | 20:40 | |
| *** souvik has quit IRC | 20:41 | |
| *** rkukura has joined #openstack-dev | 20:41 | |
| *** seanrob_ has joined #openstack-dev | 20:42 | |
| *** seanrob has quit IRC | 20:42 | |
| *** seanrob_ has quit IRC | 20:42 | |
| erkules | ahoi http://docs.openstack.org/trunk/openstack-ha/content/ch-intro.html isn't that quite to heavy on pacemaker/DRBD as HA solution? | 20:42 |
| *** seanrob has joined #openstack-dev | 20:42 | |
| *** mkollaro has quit IRC | 20:43 | |
| *** michchap has quit IRC | 20:45 | |
| *** danwent has joined #openstack-dev | 20:46 | |
| openstackgerrit | A change was merged to openstack/nova: correctly set iface-id in vmware driver https://review.openstack.org/30545 | 20:47 |
| *** bknudson has joined #openstack-dev | 20:47 | |
| *** mtreinish has quit IRC | 20:48 | |
| *** kagan has joined #openstack-dev | 20:51 | |
| *** seanrob has quit IRC | 20:53 | |
| *** seanrob has joined #openstack-dev | 20:54 | |
| *** spzala has quit IRC | 20:55 | |
| *** neelashah has quit IRC | 20:57 | |
| *** tmclaugh[work] has quit IRC | 20:58 | |
| *** jbresnah has joined #openstack-dev | 20:58 | |
| *** seanrob has quit IRC | 20:58 | |
| *** rkukura has quit IRC | 20:59 | |
| *** rkukura has joined #openstack-dev | 21:00 | |
| *** zyluo has quit IRC | 21:01 | |
| *** souvik1 has quit IRC | 21:01 | |
| *** seanrob has joined #openstack-dev | 21:02 | |
| *** souvik has joined #openstack-dev | 21:02 | |
| *** dolphm has quit IRC | 21:03 | |
| *** rkukura has quit IRC | 21:03 | |
| *** rkukura has joined #openstack-dev | 21:03 | |
| *** tmclaugh[work] has joined #openstack-dev | 21:04 | |
| *** sdake has quit IRC | 21:05 | |
| *** tmclaugh[work] has quit IRC | 21:05 | |
| *** sungju has quit IRC | 21:05 | |
| *** rnirmal has quit IRC | 21:06 | |
| *** sdake has joined #openstack-dev | 21:08 | |
| *** sdake has joined #openstack-dev | 21:08 | |
| *** morgank has quit IRC | 21:09 | |
| *** michchap has joined #openstack-dev | 21:11 | |
| *** timello has quit IRC | 21:11 | |
| *** jbresnah has quit IRC | 21:12 | |
| *** neelashah1 has joined #openstack-dev | 21:12 | |
| bknudson | what do you think about changing Keystone so that admin token doesn't have a default? | 21:12 |
| *** abhisri has quit IRC | 21:14 | |
| *** anniec has joined #openstack-dev | 21:15 | |
| *** souvik has quit IRC | 21:17 | |
| *** souvik1 has joined #openstack-dev | 21:17 | |
| *** rkukura has quit IRC | 21:17 | |
| *** dhellmann is now known as dhellmann-away | 21:17 | |
| *** flaper87 has quit IRC | 21:18 | |
| *** marun has quit IRC | 21:18 | |
| *** marun has joined #openstack-dev | 21:19 | |
| *** mrunge has quit IRC | 21:19 | |
| *** esp has joined #openstack-dev | 21:20 | |
| *** esp has left #openstack-dev | 21:20 | |
| *** devoid has quit IRC | 21:20 | |
| *** devoid has joined #openstack-dev | 21:21 | |
| *** stevemar has quit IRC | 21:21 | |
| *** devoid has quit IRC | 21:21 | |
| *** devoid has joined #openstack-dev | 21:21 | |
| *** markmcclain has quit IRC | 21:21 | |
| *** michchap has quit IRC | 21:22 | |
| *** devoid1 has joined #openstack-dev | 21:23 | |
| *** devoid has quit IRC | 21:23 | |
| *** litong has quit IRC | 21:24 | |
| openstackgerrit | A change was merged to openstack/python-swiftclient: Changed the call to set_tunnel to work in python 2.6 or python 2.7 since its name changed between versions https://review.openstack.org/32190 | 21:25 |
| *** souvik has joined #openstack-dev | 21:25 | |
| *** souvik1 has quit IRC | 21:25 | |
| *** markmcclain has joined #openstack-dev | 21:26 | |
| *** jayg is now known as jayg|g0n3 | 21:28 | |
| *** jclift has quit IRC | 21:28 | |
| *** timello has joined #openstack-dev | 21:29 | |
| *** radsy has joined #openstack-dev | 21:29 | |
| *** lorin1 has quit IRC | 21:30 | |
| *** atiwari has joined #openstack-dev | 21:32 | |
| *** jasondotstar has quit IRC | 21:36 | |
| openstackgerrit | A change was merged to openstack/oslo-incubator: Remove explicit distribute depend. https://review.openstack.org/32610 | 21:39 |
| openstackgerrit | A change was merged to openstack/oslo-incubator: Add can_send_version() to RpcProxy. https://review.openstack.org/32474 | 21:40 |
| openstackgerrit | A change was merged to openstack/oslo-incubator: zmq: remove someone unused code from ZmqClient https://review.openstack.org/31942 | 21:41 |
| openstackgerrit | A change was merged to openstack/oslo-incubator: Remove the amqp_rpc_single_reply_queue option from Havana https://review.openstack.org/31697 | 21:43 |
| *** atiwari has quit IRC | 21:43 | |
| *** eharney has quit IRC | 21:44 | |
| *** gyee has joined #openstack-dev | 21:44 | |
| *** HenryG has joined #openstack-dev | 21:45 | |
| *** rharwood has quit IRC | 21:45 | |
| openstackgerrit | A change was merged to openstack/nova: Imported Translations from Transifex https://review.openstack.org/32632 | 21:45 |
| openstackgerrit | A change was merged to openstack/cinder: Fix LVM logging error. https://review.openstack.org/32529 | 21:45 |
| *** ilyashakhat_ has quit IRC | 21:46 | |
| *** maoy has quit IRC | 21:47 | |
| *** ilyashakhat has joined #openstack-dev | 21:48 | |
| *** kagan has quit IRC | 21:48 | |
| *** afazekas has quit IRC | 21:49 | |
| *** michchap has joined #openstack-dev | 21:49 | |
| *** jkordish_ has joined #openstack-dev | 21:49 | |
| *** esp has joined #openstack-dev | 21:50 | |
| *** jkordish has quit IRC | 21:51 | |
| *** kagan has joined #openstack-dev | 21:51 | |
| *** morazi has quit IRC | 21:52 | |
| *** lbragstad has quit IRC | 21:53 | |
| *** davidkranz has quit IRC | 21:56 | |
| *** kagan has quit IRC | 21:59 | |
| *** jergerber has quit IRC | 21:59 | |
| *** esp has left #openstack-dev | 22:01 | |
| *** rackerjoe has left #openstack-dev | 22:01 | |
| *** michchap has quit IRC | 22:01 | |
| *** kbringard has quit IRC | 22:02 | |
| openstackgerrit | A change was merged to openstack/python-swiftclient: Add -l and --lh switches to swift 'list' command https://review.openstack.org/30258 | 22:03 |
| *** sungju has joined #openstack-dev | 22:03 | |
| *** markvoelker has quit IRC | 22:04 | |
| *** kashyap has quit IRC | 22:04 | |
| *** markvoelker has joined #openstack-dev | 22:05 | |
| *** adalbas has quit IRC | 22:05 | |
| *** marun_ has joined #openstack-dev | 22:06 | |
| *** kashyap has joined #openstack-dev | 22:07 | |
| *** jecarey has quit IRC | 22:07 | |
| *** anniec has quit IRC | 22:07 | |
| *** marun_ has quit IRC | 22:07 | |
| *** alexpilotti has joined #openstack-dev | 22:07 | |
| *** marun has quit IRC | 22:07 | |
| *** Tross has quit IRC | 22:08 | |
| *** ladquin_brb is now known as ladquin | 22:08 | |
| *** marun has joined #openstack-dev | 22:08 | |
| *** JordanRinke has quit IRC | 22:10 | |
| *** JordanRinke has joined #openstack-dev | 22:10 | |
| *** SergeyLukjanov has quit IRC | 22:12 | |
| *** alexpilotti has quit IRC | 22:13 | |
| *** prad_ has quit IRC | 22:14 | |
| *** matiu has joined #openstack-dev | 22:19 | |
| *** galstrom is now known as galstrom_zzz | 22:21 | |
| *** markvoelker has quit IRC | 22:22 | |
| *** dspano has quit IRC | 22:22 | |
| *** anniec has joined #openstack-dev | 22:23 | |
| *** neelashah1 has quit IRC | 22:24 | |
| *** mkollaro has joined #openstack-dev | 22:25 | |
| *** jergerber has joined #openstack-dev | 22:27 | |
| *** michchap has joined #openstack-dev | 22:28 | |
| *** jkordish_ is now known as jkordish | 22:28 | |
| *** aelkikhia has left #openstack-dev | 22:29 | |
| *** cp16net is now known as cp16net|away | 22:32 | |
| *** cp16net|away is now known as cp16net | 22:32 | |
| *** topol has joined #openstack-dev | 22:33 | |
| *** amerine has quit IRC | 22:33 | |
| *** changbl has quit IRC | 22:33 | |
| openstackgerrit | A change was merged to openstack-dev/devstack: Correct InvalidInstanceID.NotFound test. https://review.openstack.org/32346 | 22:34 |
| *** vartom110 has quit IRC | 22:34 | |
| *** giulivo has quit IRC | 22:35 | |
| *** amerine has joined #openstack-dev | 22:36 | |
| *** jergerber has quit IRC | 22:37 | |
| *** gyee has quit IRC | 22:37 | |
| *** markvoelker has joined #openstack-dev | 22:37 | |
| *** gyee has joined #openstack-dev | 22:38 | |
| *** riskable has quit IRC | 22:39 | |
| openstackgerrit | A change was merged to openstack-dev/devstack: Decrease the tempest build interval https://review.openstack.org/28660 | 22:40 |
| *** mlavalle has quit IRC | 22:40 | |
| *** mdomsch has joined #openstack-dev | 22:40 | |
| *** michchap has quit IRC | 22:40 | |
| *** riskable has joined #openstack-dev | 22:40 | |
| *** pabelanger_ has quit IRC | 22:41 | |
| *** mkollaro has quit IRC | 22:41 | |
| *** pabelanger has joined #openstack-dev | 22:41 | |
| *** santanel has quit IRC | 22:41 | |
| *** doude_ has quit IRC | 22:41 | |
| *** datsun180b has quit IRC | 22:42 | |
| *** vipul has quit IRC | 22:42 | |
| openstackgerrit | A change was merged to openstack/nova: Add deleted flag to NovaObject base https://review.openstack.org/31897 | 22:42 |
| *** akscram has quit IRC | 22:42 | |
| *** vipul has joined #openstack-dev | 22:43 | |
| *** derekh has quit IRC | 22:44 | |
| *** terryh has joined #openstack-dev | 22:44 | |
| *** akscram has joined #openstack-dev | 22:44 | |
| *** santanel has joined #openstack-dev | 22:45 | |
| openstackgerrit | A change was merged to openstack/python-novaclient: Remove explicit distribute depend. https://review.openstack.org/32623 | 22:46 |
| *** FunnyLookinHat has quit IRC | 22:46 | |
| openstackgerrit | A change was merged to openstack/nova: Refactors get_instance_security_groups to only use instance_uuid https://review.openstack.org/26441 | 22:46 |
| *** andrewbogott is now known as andrewbogott_afk | 22:46 | |
| openstackgerrit | A change was merged to openstack/cinder: Add policy checking for transfer create/accept. https://review.openstack.org/31857 | 22:46 |
| openstackgerrit | A change was merged to openstack/swift: Add non-slash delimiter tests https://review.openstack.org/31700 | 22:46 |
| openstackgerrit | A change was merged to openstack/swift: Tempauth - remove change made to HTTP_X_AUTH_TOKEN https://review.openstack.org/32370 | 22:46 |
| *** pmathews has quit IRC | 22:47 | |
| openstackgerrit | A change was merged to openstack/swift: Use threadpools in the object server for performance. https://review.openstack.org/27126 | 22:47 |
| *** morganfainberg has left #openstack-dev | 22:47 | |
| openstackgerrit | A change was merged to openstack/nova: Fix assumed port has port_security_enabled https://review.openstack.org/32288 | 22:47 |
| *** sushils has joined #openstack-dev | 22:47 | |
| *** jpich has quit IRC | 22:47 | |
| *** andrewbogott_afk is now known as andrewbogott | 22:47 | |
| *** doude has joined #openstack-dev | 22:48 | |
| openstackgerrit | A change was merged to openstack/swift: Add parallelism to object expirer daemon. https://review.openstack.org/27966 | 22:48 |
| *** atiwari has joined #openstack-dev | 22:48 | |
| *** jimfehlig has quit IRC | 22:49 | |
| openstackgerrit | A change was merged to openstack/nova: Add base mixin class for object lists https://review.openstack.org/31842 | 22:49 |
| *** newbie-999 has joined #openstack-dev | 22:49 | |
| *** carl_baldwin has quit IRC | 22:50 | |
| *** markmcclain has quit IRC | 22:51 | |
| *** sushils has quit IRC | 22:51 | |
| newbie-999 | * Question about running a dashboard off a devstack setup please. My dashboard has 'Action' column for Instances and Volumes. Under this column, the 'More' button is doing nothing. No drop down! | 22:52 |
| *** sride has quit IRC | 22:52 | |
| *** egallen has quit IRC | 22:52 | |
| *** HenryG_ has joined #openstack-dev | 22:52 | |
| *** souvik has quit IRC | 22:53 | |
| *** henrynash has quit IRC | 22:53 | |
| *** sushils has joined #openstack-dev | 22:53 | |
| *** HenryG has quit IRC | 22:56 | |
| *** terryh has quit IRC | 22:56 | |
| *** cmagina has quit IRC | 22:56 | |
| *** electrichead has quit IRC | 22:56 | |
| *** cmagina has joined #openstack-dev | 22:57 | |
| newbie-999 | * This is off the latest code. | 22:57 |
| *** mkollaro has joined #openstack-dev | 23:00 | |
| *** cmagina has quit IRC | 23:05 | |
| *** michchap has joined #openstack-dev | 23:06 | |
| *** cmagina has joined #openstack-dev | 23:08 | |
| *** danwent has quit IRC | 23:08 | |
| *** Ruetobas has joined #openstack-dev | 23:11 | |
| *** huats_ has joined #openstack-dev | 23:13 | |
| *** huats_ has quit IRC | 23:13 | |
| *** huats_ has joined #openstack-dev | 23:13 | |
| *** huats has quit IRC | 23:14 | |
| *** savid has quit IRC | 23:15 | |
| *** retr0h_ is now known as retr0h | 23:18 | |
| *** michchap has quit IRC | 23:18 | |
| *** jamielennox|away is now known as jamielennox | 23:21 | |
| *** lloydde has quit IRC | 23:21 | |
| *** tzumainn has quit IRC | 23:22 | |
| *** varora200 has quit IRC | 23:23 | |
| *** rcleere has quit IRC | 23:28 | |
| *** danwent has joined #openstack-dev | 23:28 | |
| *** rcleere has joined #openstack-dev | 23:29 | |
| *** mdomsch has quit IRC | 23:29 | |
| *** krtaylor has quit IRC | 23:29 | |
| *** msmedved has quit IRC | 23:31 | |
| *** akscram has quit IRC | 23:35 | |
| *** sride has joined #openstack-dev | 23:36 | |
| *** akscram has joined #openstack-dev | 23:36 | |
| *** mrodden has quit IRC | 23:36 | |
| *** newbie-999 has quit IRC | 23:38 | |
| *** asalkeld is now known as gloomy | 23:39 | |
| *** otherwiseguy has quit IRC | 23:39 | |
| *** gloomy is now known as angus | 23:39 | |
| *** devoid1 has quit IRC | 23:41 | |
| *** lloydde has joined #openstack-dev | 23:44 | |
| *** michchap has joined #openstack-dev | 23:44 | |
| *** michchap has quit IRC | 23:45 | |
| *** michchap has joined #openstack-dev | 23:46 | |
| *** redbeard2 has quit IRC | 23:47 | |
| *** atiwari has quit IRC | 23:56 | |
| *** mkollaro has quit IRC | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!