| fungi | but yeah, i suspect her script which sets up teh cgit config also needs to init any missing repos | 00:00 |
|---|---|---|
| fungi | should be a fairly minor addition since it's already doing stuff with the projects.yaml registry | 00:00 |
| mordred | fungi: any idea where that script is? | 00:01 |
| fungi | jeepyb | 00:01 |
| mordred | fungi: k. also - any idea where it's called in puppet? | 00:01 |
| fungi | create_cgitrepos.py | 00:01 |
| fungi | called from modules/cgit/init.pp i think... checking | 00:02 |
| mordred | ah. I see it | 00:02 |
| mordred | neat | 00:02 |
| fungi | ahh, modules/openstack_project/manifests/git.pp | 00:03 |
| openstackgerrit | A change was merged to openstack-infra/config: Remove if module statements https://review.openstack.org/41278 | 00:07 |
| fungi | hmm, marun's gone, but i manually spun up a devstack vm and have some tempest tests running on it no problem | 00:07 |
| fungi | just following https://github.com/openstack-infra/devstack-gate#simulating-devstack-gate-tests (so presumably something we're doing in prep is causing the issue not to surface) | 00:08 |
| mordred | fungi: that's what I'm thinking | 00:10 |
| mordred | fungi: I replicated it not working when we don't run our stuff | 00:10 |
| mordred | fungi: I'm guessing something in the puppet setup is overreaching | 00:10 |
| fungi | yep. clearly some of our stuff should be in devstack ;) | 00:10 |
| mordred | hehe | 00:11 |
| openstackgerrit | Monty Taylor proposed a change to openstack-infra/jeepyb: Create repo when we create the cgit config file https://review.openstack.org/41279 | 00:11 |
| mordred | fungi, pleia2 | 00:11 |
| mordred | fungi, pleia2: ^^ there's a stab - it probably doesn't work | 00:11 |
| fungi | i wonder if my previous instructions for a puppetless workflow hits the same issues | 00:11 |
| fungi | if not, should make it easier to narrow down | 00:12 |
| * mordred also just wants to make the global venv work | 00:12 | |
| mordred | in devstack | 00:12 |
| mordred | because it should isolate most of these sorts of things | 00:12 |
| *** Ryan_Lane has joined #openstack-infra | 00:12 | |
| *** zul has joined #openstack-infra | 00:22 | |
| *** pabelanger has quit IRC | 00:26 | |
| *** pcrews has joined #openstack-infra | 00:33 | |
| *** Ryan_Lane has quit IRC | 00:35 | |
| fungi | dug out my puppetless simulation instructions and have them running now to see if they hit it | 00:37 |
| mordred | fungi: you has link? | 00:41 |
| *** sarob has quit IRC | 00:42 | |
| *** sarob has joined #openstack-infra | 00:42 | |
| fungi | mordred: https://review.openstack.org/#/c/32661/3/README.rst | 00:47 |
| fungi | is what i'm trying | 00:47 |
| *** sarob has quit IRC | 00:47 | |
| fungi | it's in devstack exercises still at the moment | 00:48 |
| mordred | hrm | 00:49 |
| koolhead17 | hoi all | 00:49 |
| fungi | mordred: it made it through exercises and has moved on to tempest | 00:52 |
| mordred | fungi: then it's working | 00:53 |
| fungi | and it's not like i really did anything too special there... updated a base precise image, installed debs of git and python-pip, used pip to install tox, then basic user stuff and devstack-vm-gate-wrap | 00:54 |
| mordred | oh. wait | 00:54 |
| mordred | you ran devstack-vm-gate-wrap | 00:54 |
| fungi | so i'm thinking it's that, yes | 00:54 |
| mordred | so, what I'm doing to test the failure case | 00:54 |
| fungi | which means it's probably not something we're doing in puppet, but rather something we're doing in d-g | 00:55 |
| mordred | is "new machine, apt-get install git, git clone devstack, (make localrc because it's on hpcloud and I have to set network range), ./stack.sh | 00:55 |
| fungi | which is preventing the issue from cropping up in the gate | 00:55 |
| *** nayward has joined #openstack-infra | 00:57 | |
| *** nayward has quit IRC | 01:02 | |
| *** Ryan_Lane has joined #openstack-infra | 01:04 | |
| mordred | fungi: I don't see anyuthing in there which makes senes to me to be a problem | 01:05 |
| fungi | on the one i built by applying our puppet config, after the full tempest run completed, successfully, i was able to unstack and stack again with no issue | 01:05 |
| *** vipul-away is now known as vipul | 01:05 | |
| fungi | going to try the same on the puppetless one in a bit | 01:06 |
| fungi | but i expect the results to be similar | 01:06 |
| *** Ryan_Lane has quit IRC | 01:17 | |
| *** zul has quit IRC | 01:21 | |
| *** pcrews has quit IRC | 01:30 | |
| *** zul has joined #openstack-infra | 01:43 | |
| *** pabelanger has joined #openstack-infra | 01:49 | |
| fungi | [ERROR] /opt/stack/devstack/lib/keystone:316 keystone did not start | 01:50 |
| fungi | so on the non-puppeted vm, even though exercises and tempest ran successfully from the wrap script, unstack and stack afterward fails | 01:50 |
| fungi | nevermind, it does not. i was looking at the wrong vm. the error above was from one i tried with no puppet and without running the wrap script first | 01:52 |
| fungi | so if we run the wrap script to initiate tests, then devstack is configured such that it can be successfully unstacked and stacked again afterward | 01:52 |
| jeblair | fungi, mordred: you're working on a new experimental thing, not fixing a production problem, right? | 01:53 |
| fungi | jeblair: no, trying to figure out why devstack has gotten itself in a state where it doesn't work on its own, but works from the gate wrap script | 01:54 |
| jeblair | fungi: ugh. ok. | 01:55 |
| fungi | continuation of the issues we saw over the past couple days where, i think, pip is installing some packages with non-world-readable permissions | 01:56 |
| jeblair | fungi, mordred: jenkins01 doesn't appear to have done anything insane, so i'm putting jenkins02 into shutdown mode in prep for upgrading gearman-plugin there | 01:58 |
| fungi | great! | 01:58 |
| *** sarob has joined #openstack-infra | 02:00 | |
| clarkb | jeblair how soon do you want to get the new pool code running? I plan to review it tomorrow | 02:02 |
| *** vipul is now known as vipul-away | 02:02 | |
| *** Ryan_Lane has joined #openstack-infra | 02:03 | |
| *** Ryan_Lane has joined #openstack-infra | 02:03 | |
| *** sarob has quit IRC | 02:04 | |
| jeblair | clarkb: a few more things need to happen first, so don't rush. in particular, zuul needs a new feature (to run a custom function at job start and completion) that needs to be written and installed first. so it'll probably be a couple days into next week at least, if not the end of the week. | 02:05 |
| *** ewindisch has joined #openstack-infra | 02:06 | |
| mordred | fungi: AWESOME | 02:07 |
| mordred | jeblair: and yeah, what fungi said | 02:07 |
| mordred | I think it's definitely a thing to figure out | 02:08 |
| *** zul has quit IRC | 02:08 | |
| mordred | jeblair: I honestly can't see what d-g might do that plain doesn't | 02:08 |
| mordred | oh | 02:08 |
| mordred | fungi: you know what? what if it's the mirror? | 02:08 |
| mordred | what if there IS something about the mirror which is different, and is causing devstack runs pointing at it to succeed and ones not pointing at it to fail | 02:09 |
| fungi | mmm, does the wrap script set it to use our pypi mirror instead of the real pypi? | 02:09 |
| * fungi rtfs | 02:09 | |
| mordred | fungi: yes | 02:09 |
| openstackgerrit | A change was merged to openstack-dev/pbr: Update requirements in integration test https://review.openstack.org/41098 | 02:12 |
| clarkb | should be easy to test that just set the mirror and run stack.sh? | 02:13 |
| mordred | yup. doing now | 02:13 |
| openstackgerrit | A change was merged to openstack-dev/pbr: Added documentation for packagers https://review.openstack.org/41104 | 02:14 |
| mordred | yay! | 02:15 |
| mordred | docs fixed | 02:15 |
| mordred | http://docs.openstack.org/developer/pbr/ | 02:15 |
| fungi | i have it running now | 02:15 |
| mordred | now shows most recent docs | 02:15 |
| fungi | should know here in a moment | 02:16 |
| fungi | ick, i think i got the machine into a state | 02:17 |
| *** erfanian has joined #openstack-infra | 02:18 | |
| fungi | okay, have it going now i think | 02:19 |
| lifeless | mordred: any comment on my confusion? | 02:23 |
| mordred | lifeless: you have so much confusion, and I've been battling devstack, can you clarify? | 02:24 |
| mordred | ;) | 02:24 |
| mordred | lifeless: also, you might like this followup patch: https://review.openstack.org/#/c/41273/ | 02:25 |
| mordred | lifeless: AND - I split things into functions for you | 02:25 |
| *** bearhands is now known as comstud | 02:28 | |
| lifeless | mordred: 10:27 < lifeless> mordred: so what about this sentence - | 02:28 |
| lifeless | 10:27 < lifeless> # Need to do these separately to not trigger setuptools | 02:28 |
| lifeless | 10:27 < lifeless> # Bug. | 02:28 |
| lifeless | 10:27 < lifeless> I still can't make head or tails of the end of it. | 02:28 |
| mordred | AH | 02:28 |
| lifeless | mordred: thats the only concern I have remaining about the patch | 02:28 |
| lifeless | I commented in a review, but there was no reply, nor change. | 02:28 |
| mordred | that. I should probably not use the word bug | 02:28 |
| mordred | as it's not a thing that's fixable/going to be fixed aiui | 02:28 |
| lifeless | mordred: possibly. Also possibly not as a proper noun :) | 02:29 |
| mordred | lemme go clarity - sorry, forgot to update that bit | 02:29 |
| lifeless | thanks! | 02:30 |
| mordred | lifeless: this is the thing where if you upgrade setuptools _with_ antyhing else, you might trigger a condition where distribute upgrades, removing its version of the setuptools code, but another package also attempts to upgrade before setuptools gets reinstalled | 02:30 |
| lifeless | mordred: so arguably thats a bug in the replacement distribute | 02:31 |
| *** ewindisch has quit IRC | 02:31 | |
| mordred | lifeless: yes | 02:31 |
| mordred | well, actually | 02:31 |
| lifeless | mordred: personally I *would* file an upstream bug, and reference that. | 02:31 |
| mordred | it's in pip's depenency processing | 02:31 |
| lifeless | potato potato | 02:31 |
| mordred | and I_believe_ clarkb filed a bug | 02:31 |
| lifeless | mordred: I'd just like either a) a bug #, or b) a short description, or ideally c) both. | 02:31 |
| *** kiall has quit IRC | 02:34 | |
| *** kiall has joined #openstack-infra | 02:37 | |
| *** mberwanger has joined #openstack-infra | 02:37 | |
| fungi | resetting the same pip/pydistutils configs to our mirror as d-g does, i still get "keystone did not start" so i'm thinking it's not our pypi mirror making the difference | 02:40 |
| fungi | but am going to try again on a fresh vm just to be sure | 02:40 |
| mordred | lifeless: https://github.com/pypa/pip/issues/988 | 02:41 |
| mordred | fungi: oh good | 02:41 |
| lifeless | mordred: I don't believe it's the same bug | 02:42 |
| mordred | lifeless: might not be | 02:42 |
| lifeless | mordred: thats scheduling not resolving, similar but different problems. | 02:42 |
| lifeless | mordred: specifically it could resolve correctly and not force adjacent scheduling for setuptools&distribute. | 02:42 |
| mordred | indeed | 02:43 |
| mordred | of course, I think it would have a very hard time | 02:43 |
| mordred | given that dstufft tells me best practice should be to not declare a setuptools dependency in a package | 02:43 |
| dstufft | mostly because it's broken to do so | 02:44 |
| mordred | so I kinda think that pip should have special handling of setuptools depends | 02:44 |
| mordred | dstufft: well, the thing is | 02:44 |
| mordred | if you don't | 02:44 |
| dstufft | pip 1.4 does have special handling of the setuptools/distribute depends | 02:44 |
| dstufft | it moves it to the uh, end (or is it begining?) of the installation set | 02:44 |
| mordred | great! so the setuptools upgrade break goes away with pip 1.4? | 02:44 |
| dstufft | That's what I'm told, clarkb has another case that is exposed with older versions of python-mysql that I didn't dig into yet | 02:45 |
| dstufft | But that existed prior to the setuotools/distribute 0.7 | 02:45 |
| mordred | well, they declare an explicit depend on distribute | 02:45 |
| mordred | as does d2to1 (although we stopped consuming it, mainly for that reason) | 02:45 |
| dstufft | but the issues that came about because of 0.7+ should be resolved in pip 1.4+ | 02:46 |
| dstufft | from what I'm told at least, I didn't fix that problem | 02:46 |
| * mordred thinks about how he could test this... | 02:46 | |
| dstufft | mordred: fwiw it was explicitly a problem with our recursive upgrades, distribute 0.7 depending on setuptools, and distribute not having a setuptools directory itself | 02:49 |
| mordred | dstufft: https://review.openstack.org/41281 | 02:50 |
| mordred | dstufft: I just sent that in to devstack, which removes the explicit setuptools upgrade, and instead only upgades pip | 02:50 |
| dstufft | which broke our hack to make upgrading setuptools work (by in the subshell that calls setup.py install, changing cwd to the build dir so it could import the setuptools from inside the package) | 02:50 |
| dstufft | mordred: ok | 02:52 |
| dstufft | mordred: hopefully it works and I don't look silly :) | 02:52 |
| mordred | dstufft: if it works, it will invovle simplication of things, which will make me happy | 02:53 |
| mordred | dstufft: if it doesn't work, you've indicated that you think it _should_ which means maybe we could get a 1.4.2 out amending the logic to fix :) | 02:53 |
| mordred | I'll be happy either way | 02:53 |
| *** erfanian has quit IRC | 02:53 | |
| openstackgerrit | Monty Taylor proposed a change to openstack-infra/jeepyb: Remove spurious uses of output return values https://review.openstack.org/41273 | 02:53 |
| openstackgerrit | Monty Taylor proposed a change to openstack-infra/jeepyb: Start building wheels in the mirror https://review.openstack.org/41215 | 02:53 |
| mordred | lifeless: ^^ updated comment to be verbose - did not include bug number. also upgraded pip first | 02:54 |
| mordred | because either we need to upgade setuptools by itself to avoid the bug, or we need to ensure our virtualenv has the latest pip to avoid the bug :) | 02:54 |
| mordred | but latest pip seems long-term to be the stronger over all solution | 02:55 |
| dstufft | mordred: you can just download virtualenv-X.Y.tar.gz, unpack it and create your virtualenv like python virtualenv-X.Y/virtualenv.py --options | 02:55 |
| mordred | dstufft: that's actually probably more work in this context - but thanks, good to know | 02:55 |
| dstufft | and get the version of pip and setuptools bundled with that virtualenv | 02:55 |
| dstufft | okies | 02:55 |
| dstufft | mordred: only mentioned it because it might be nice to be completely indepedent from system packages | 02:56 |
| * mordred may pull that trick out of sleeve later... | 02:56 | |
| mordred | dstufft: yes. indeed | 02:56 |
| mordred | I will strongly consider making that how devstack gets its virtualenv | 02:56 |
| mordred | if we can sell folks on global virtualenv | 02:56 |
| mordred | gosh - if only we could get libvirt in virtualenv... | 02:57 |
| *** nayward has joined #openstack-infra | 02:57 | |
| mordred | dstufft: w00t! it may be working | 02:58 |
| dstufft | mordred: :D | 02:59 |
| * dstufft may not look silly | 02:59 | |
| dstufft | at least not for this reason! | 02:59 |
| *** sarob has joined #openstack-infra | 03:00 | |
| *** nayward has quit IRC | 03:02 | |
| *** dkliban has joined #openstack-infra | 03:03 | |
| *** sarob has quit IRC | 03:05 | |
| *** Ryan_Lane has quit IRC | 03:09 | |
| *** Ryan_Lane has joined #openstack-infra | 03:20 | |
| *** Ryan_Lane has quit IRC | 03:25 | |
| openstackgerrit | Alessandro Pilotti proposed a change to openstack-dev/pbr: Fixes issue with command escaping on Windows https://review.openstack.org/41283 | 03:30 |
| openstackgerrit | Alessandro Pilotti proposed a change to openstack-dev/pbr: Fixes issue with command escaping on Windows https://review.openstack.org/41283 | 03:31 |
| openstackgerrit | Alessandro Pilotti proposed a change to openstack-dev/pbr: Fixes issue with command escaping on Windows https://review.openstack.org/41283 | 03:32 |
| *** pcrews has joined #openstack-infra | 03:36 | |
| *** Ryan_Lane has joined #openstack-infra | 03:41 | |
| *** dkehn has quit IRC | 03:47 | |
| *** dkehn has joined #openstack-infra | 03:47 | |
| jeblair | mordred, fungi: i've restarted jenkins02 with new gearman-plugin | 03:54 |
| fungi | awesome. no signs of problems with it sounds like? | 03:54 |
| jeblair | fungi: doesn't look like it; jenkins01 seems to be behaving | 03:55 |
| *** sarob has joined #openstack-infra | 04:01 | |
| mordred | jeblair: w00t | 04:05 |
| mordred | jeblair: rolling upgrades ftw | 04:05 |
| *** sarob has quit IRC | 04:06 | |
| jeblair | ++ | 04:07 |
| fungi | mordred: !!! confirmed with a fresh vm that overriding pip/pydistutils config to point at our pypi mirror does in fact make it work. ran the following... http://paste.openstack.org/show/43829/ | 04:08 |
| mordred | fungi: oh god | 04:09 |
| fungi | so, um... i wonder if there's something different about the packages in question on our mirror vs on pypi proper? (prettytable, what else?) | 04:09 |
| mordred | fungi: that's just absolutely great | 04:10 |
| fungi | someone repeating that experiment to confirm would be good, just in case i did something stupid | 04:11 |
| fungi | in the meantime, i'm pulling packages from our mirror and pypi to compare | 04:11 |
| mordred | fungi: do you still have your working one in? | 04:11 |
| mordred | fungi: try running "pip list" on both | 04:11 |
| mordred | or, rather, the one that didn't work | 04:11 |
| fungi | oh, no but i can make one which doesn't work and compare them | 04:12 |
| mordred | hrm. pip list blows up | 04:12 |
| mordred | pip freeze works | 04:12 |
| mordred | dstufft: is that someone you want more info from me on? | 04:12 |
| dstufft | define blow up | 04:13 |
| mordred | dstufft: http://paste.openstack.org/show/43830/ | 04:13 |
| mordred | dstufft: pip freeze operates to completion | 04:14 |
| mordred | fungi: I'm _guessing_ that our mirror has unwittingly expressed a version pin that should (clearly) actually be in requirements.txt | 04:15 |
| dstufft | mordred: Can you file a bug? That shouldn't happen. Ideally with the pip freeze and pip list output | 04:15 |
| mordred | fungi: figuring out what that pin is (a) and why we're skewed (b) I think are both worthwhile | 04:15 |
| mordred | dstufft: totaly! | 04:15 |
| * dstufft afks | 04:16 | |
| fungi | mordred: quite possibly because our means of caching by installing is failing to add the new version when installation goes wrong? | 04:16 |
| fungi | mordred: but yeah, i'll diff pip freeze between working and broken | 04:16 |
| mordred | dstufft: https://github.com/pypa/pip/issues/1093 | 04:17 |
| *** sarob has joined #openstack-infra | 04:20 | |
| *** Ryan_Lane has quit IRC | 04:22 | |
| *** mberwanger has quit IRC | 04:23 | |
| openstackgerrit | Alessandro Pilotti proposed a change to openstack-dev/pbr: Fixes issue with command escaping on Windows https://review.openstack.org/41283 | 04:26 |
| *** mrodden1 has quit IRC | 04:27 | |
| *** locke105 has quit IRC | 04:27 | |
| *** sarob has quit IRC | 04:29 | |
| *** sarob has joined #openstack-infra | 04:29 | |
| mordred | fungi: if you feel like looking at an easy pbr bug, ^^ is blocking our windows/hyper-v folks | 04:29 |
| *** oldben has quit IRC | 04:30 | |
| *** bnemec has joined #openstack-infra | 04:30 | |
| *** locke105 has joined #openstack-infra | 04:30 | |
| *** mrodden has joined #openstack-infra | 04:31 | |
| *** sarob has quit IRC | 04:34 | |
| mordred | ok. I think I may try to go to the sleeping hole | 04:36 |
| fungi | mordred: diffing pip freeze output between failing and working runs was unhelpful. the failing one is simply missing extras, fixtures, python-mimeparse, python-subunit, tempest-master, testrepository, testresources and testtools | 04:36 |
| fungi | and yeah, i probably need to think about sleep as well | 04:37 |
| fungi | though pip freeze is also struck by the permissions issue on the failing run... IOError: [Errno 13] Permission denied: '/usr/local/lib/python2.7/dist-packages/p | 04:38 |
| fungi | rettytable-0.7.2.egg-info/top_level.txt' | 04:38 |
| *** zul has joined #openstack-infra | 04:41 | |
| *** pcrews has quit IRC | 04:43 | |
| fungi | worse still, i've downloaded the .tar.bz2, .tar.gz and .zip copies of prettytable from pypi and from our mirror and their checksums match | 04:47 |
| fungi | think i'm done for the night. i'll continue seeing if i can come up with something tomorrow | 04:55 |
| *** nayward has joined #openstack-infra | 04:58 | |
| *** nayward has quit IRC | 05:03 | |
| *** sarob has joined #openstack-infra | 05:10 | |
| *** ewindisch has joined #openstack-infra | 05:19 | |
| *** ewindisch has quit IRC | 05:32 | |
| *** sarob has quit IRC | 05:42 | |
| *** sarob has joined #openstack-infra | 05:42 | |
| *** sarob has quit IRC | 05:47 | |
| *** Ryan_Lane has joined #openstack-infra | 05:48 | |
| *** Ryan_Lane has joined #openstack-infra | 05:48 | |
| *** Ryan_Lane has quit IRC | 05:55 | |
| *** nayward has joined #openstack-infra | 06:09 | |
| *** nayward has quit IRC | 06:15 | |
| *** nayward has joined #openstack-infra | 06:35 | |
| *** nayward has quit IRC | 06:40 | |
| *** sarob has joined #openstack-infra | 06:51 | |
| *** nayward has joined #openstack-infra | 06:54 | |
| *** sarob has quit IRC | 06:56 | |
| *** mikal has quit IRC | 07:07 | |
| *** mikal has joined #openstack-infra | 07:08 | |
| *** Ryan_Lane has joined #openstack-infra | 07:15 | |
| *** yolanda has joined #openstack-infra | 07:26 | |
| *** koobs has quit IRC | 07:35 | |
| *** sarob has joined #openstack-infra | 07:52 | |
| *** sarob has quit IRC | 07:55 | |
| *** yolanda has quit IRC | 07:59 | |
| *** yolanda has joined #openstack-infra | 08:12 | |
| *** yolanda has quit IRC | 08:18 | |
| *** nayward has quit IRC | 08:19 | |
| *** sarob has joined #openstack-infra | 08:55 | |
| *** nicedice_ has quit IRC | 08:56 | |
| *** Ryan_Lane has quit IRC | 08:56 | |
| *** sarob has quit IRC | 09:00 | |
| *** Ryan_Lane has joined #openstack-infra | 09:12 | |
| *** nicedice_ has joined #openstack-infra | 09:49 | |
| *** Ryan_Lane has quit IRC | 09:52 | |
| *** sarob has joined #openstack-infra | 09:57 | |
| *** sarob has quit IRC | 10:04 | |
| *** sdake_ has quit IRC | 10:44 | |
| *** sarob has joined #openstack-infra | 11:01 | |
| *** sarob has quit IRC | 11:05 | |
| openstackgerrit | Sean Dague proposed a change to openstack-infra/config: unified styling on progress bars https://review.openstack.org/41299 | 11:15 |
| *** SergeyLukjanov has joined #openstack-infra | 11:19 | |
| *** SergeyLukjanov has quit IRC | 11:24 | |
| *** mikal has quit IRC | 11:38 | |
| *** mikal has joined #openstack-infra | 11:40 | |
| *** sarob has joined #openstack-infra | 12:01 | |
| *** sarob has quit IRC | 12:06 | |
| *** mikal has quit IRC | 12:11 | |
| *** mikal has joined #openstack-infra | 12:12 | |
| *** dkliban has quit IRC | 12:38 | |
| *** mjfork has joined #openstack-infra | 12:49 | |
| *** dripton_ has joined #openstack-infra | 12:52 | |
| *** dripton has quit IRC | 12:53 | |
| *** dkliban has joined #openstack-infra | 12:57 | |
| *** mjfork has quit IRC | 13:00 | |
| *** mjfork has joined #openstack-infra | 13:03 | |
| *** krtaylor_ has quit IRC | 13:08 | |
| *** boris-42 has joined #openstack-infra | 13:11 | |
| *** dkliban has quit IRC | 13:13 | |
| *** sgviking has quit IRC | 13:27 | |
| mordred | fungi: the prettytable permissions issues are something marun and dtroyer are working on too | 13:28 |
| mordred | so maybe _something_ about how we do the puppet nodes is preventing the pretty table permissions issues from arising | 13:28 |
| mordred | morning sdague | 13:29 |
| *** sarob has joined #openstack-infra | 13:32 | |
| *** sarob has quit IRC | 13:36 | |
| zaro | mordred: good morning. | 13:58 |
| mordred | morning zaro ! | 13:58 |
| openstackgerrit | A change was merged to openstack-infra/jenkins-job-builder: Improved the gerrit trigger https://review.openstack.org/40644 | 13:59 |
| zaro | mordred: you have time to review today? https://review.openstack.org/#/c/40455/ | 14:00 |
| mordred | looking | 14:00 |
| openstackgerrit | A change was merged to openstack-infra/config: deploy jenkins plugin pom.xml file. https://review.openstack.org/40455 | 14:05 |
| zaro | mordred: thanks. Wow! builds have become super snappy since the additional jenkins masters. | 14:06 |
| mordred | zaro: yes they have! | 14:10 |
| *** sarob has joined #openstack-infra | 14:21 | |
| openstackgerrit | Khai Do proposed a change to openstack-infra/gearman-plugin: moved python examples to jenkins wiki https://review.openstack.org/41307 | 14:22 |
| *** sarob has quit IRC | 14:27 | |
| *** thomasm has joined #openstack-infra | 14:29 | |
| openstackgerrit | Monty Taylor proposed a change to openstack-dev/pbr: Fix issue with command escaping on Windows https://review.openstack.org/41283 | 14:31 |
| fungi | mordred: i don't think it's puppet fixing the problem. as i pointed out last night, simply configuring pip/pydistutils to use pypi.openstack.org instead of pypi.python.org causes those permissions issues not to arise | 14:32 |
| fungi | and yet, the package versions and checksums are identical | 14:33 |
| *** Ryan_Lane has joined #openstack-infra | 14:33 | |
| mordred | fungi: oh right. but, you know, if the versions aren't different | 14:33 |
| mordred | then how are we getting different behavior? | 14:33 |
| fungi | at least the versions of packages it got far enough to install were all the same. the only checksums i compared were for prettytable itself though | 14:34 |
| mordred | fungi: just out of curiosity, can you check the version of pip on each machine? | 14:34 |
| fungi | both of them say "pip 1.4 from /usr/local/lib/python2.7/dist-packages/pip-1.4-py2.7.egg (python 2.7)" | 14:35 |
| fungi | though i want to try building them again, dumping the devstack setup spam to a file and comparing some context around where prettytable is being installed to see if there's any difference | 14:36 |
| fungi | they also both have distribute==0.6.24dev-r0 and no setuptools (according to pip freeze) | 14:38 |
| fungi | i suppose i should script something to retrieve the entire output of pip freeze from our mirror and upstream and see if there's any invalid checksums | 14:42 |
| *** thomasm has quit IRC | 14:43 | |
| fungi | oh, though i'd need to filter out the packages which are being reported by pip freeze but actually installed from debs | 14:44 |
| *** sarob has joined #openstack-infra | 14:46 | |
| mordred | hrm | 14:47 |
| mordred | fungi: btw - it's driving me batty getting emails from OpenStack Hudson (sometimes I just finally can't take it any longer) | 14:47 |
| mordred | fungi: so I'd like to change the display name at least (I know changing the actual account name is WAY more involved) | 14:47 |
| mordred | jeblair: ^^ | 14:48 |
| fungi | mordred: should be fine | 14:48 |
| fungi | mordred: changing the display name should be fine i mean | 14:48 |
| mordred | fungi: I was thinking it should maybe be "OpenStack Infra" ? | 14:48 |
| mordred | or somethig - Jenkins is equally wrong | 14:48 |
| fungi | yeah, it shouldn't be something specific and wrong | 14:49 |
| fungi | i'm okay with it being something specific and accurate | 14:49 |
| fungi | however, specific and accurate eventually drifts until it's specific and wrong again | 14:49 |
| fungi | so general helps prevent that, at the risk of being a bit vague | 14:49 |
| openstackgerrit | Alessandro Pilotti proposed a change to openstack-dev/pbr: Fixes issue with command escaping on Windows https://review.openstack.org/41283 | 14:51 |
| mordred | fungi: so, Infra general enough? Or spell it out Infrastructure? | 14:52 |
| jeblair | mordred: i'd go with infra (since we want to rename the project to that too) | 14:53 |
| fungi | yeah, agreed | 14:53 |
| mordred | k | 14:53 |
| clarkb | ++ | 14:54 |
| mordred | jeblair, fungi: done | 14:54 |
| fungi | yay! | 14:54 |
| jeblair | i unstuck puppet on jenkins.o.o | 14:54 |
| jeblair | (its jenkins jobs were out of date) | 14:55 |
| jeblair | fungi, mordred: err: /Stage[main]/Openstack_project::Base/Package[virtualenv]/ensure: change from 1.10 to 1.10.1 failed: Could not update: Execution of '/usr/bin/pip install -q --upgrade virtualenv' returned 1: Command python setup.py egg_info failed with error code 1 | 14:56 |
| jeblair | fungi, mordred: what's the current fix for that? | 14:56 |
| mordred | sigh | 14:56 |
| mordred | jeblair: well - I have learned some things since last we spoke | 14:56 |
| mordred | jeblair: on this topic | 14:56 |
| mordred | namely, apparently we do _not_ need to take pains to upgrade the system setuptools before everything else | 14:57 |
| mordred | as long as we upgrade pip to >=1.4 (1.4 has special case logic to handle the breakage issues) | 14:57 |
| mordred | jeblair: I have not translated that information in to a new overall plan for our puppet modules | 14:57 |
| mordred | but I shall endeavor to do so | 14:57 |
| clarkb | you don't want pip to ensure latest pip | 14:57 |
| clarkb | because pip freeze does not report pip which was marked an invalid bug by upstream... | 14:58 |
| mordred | yah | 14:58 |
| zaro | is it faux pas to approve your own patches, even though it's trivial? | 15:01 |
| clarkb | maybe deb/rpm packages for pip is all we need? | 15:01 |
| mordred | yes. I agree | 15:02 |
| jeblair | zaro: usually; we sometimes make exceptions when we're fixing operational problems. | 15:02 |
| mordred | yeah. what jeblair said- and even then I try to keep it to only things that are super simple - if it's a complex operational fix, I want someone to look at it | 15:03 |
| mordred | clarkb: the problem is, no one upstream has made a pip 1.4 pakckage yet | 15:04 |
| mordred | clarkb: and once I start thinking down that road, my brain wants to know how we're going to manage it in an ongoing manner | 15:04 |
| sdague | mordred: morning | 15:06 |
| openstackgerrit | A change was merged to openstack/requirements: update mock minimum to 1.0 https://review.openstack.org/41168 | 15:06 |
| openstackgerrit | A change was merged to openstack/requirements: Always specify a lower bound on packages https://review.openstack.org/37911 | 15:06 |
| sdague | jeblair: was playing with status bar styling in html5 this morning, second patch up for non-native styling - (looks like this - http://img849.imageshack.us/img849/6774/r9se.png) | 15:09 |
| jeblair | sdague: nice! i think the color should probably be #e2ecef (the background for the bar at the top of the change box) | 15:12 |
| sdague | jeblair: so if you make it e2ecef you can't really see it | 15:13 |
| sdague | I did start there | 15:13 |
| sdague | but it's actually too subtle | 15:13 |
| jeblair | sdague: do you know of a way to do leaders in html? i'm having trouble following from the short jobs (gate-pbr-docs) over to the results? | 15:13 |
| jeblair | how about #95c7db? | 15:14 |
| sdague | jeblair: leaders? | 15:14 |
| jeblair | sdague: like......this | 15:14 |
| jeblair | sdague: they lead your eye | 15:14 |
| sdague | ah, yeh I'm not sure | 15:15 |
| jeblair | sdague: with a really big gap like that, you actually have to stop tracking while reading and lose focus in order jump from one side to the other | 15:15 |
| sdague | jeblair: fair, though from a interface perspective I think most folks are actually going to that page to find their job, scan for fails, and look back only if there is one | 15:16 |
| jeblair | sdague: interesting thought -- design for reading backwards. :) | 15:17 |
| sdague | yeh, that's why I wanted the status floated and lined up, because that's typically what I'm trying to pull out | 15:18 |
| sdague | anyway ux agreement is always hard, everyone comes at it with a different pov. :) | 15:18 |
| jeblair | sdague: that's very compelling, and makes the leaders (or something equivalent, table lines, bg color, etc) seem less necessary, though i suspect (i've never tried myself) they can be useful for reading backwards too. at any rate, i don't think that needs to block this. | 15:20 |
| jeblair | sdague: i do think the color needs to be compatible with the blues there though; did you try #95c7db? | 15:21 |
| sdague | I did I found it really overwhelmingly dark, so that's all your eye sees | 15:21 |
| sdague | let me do a couple a screen shots to show | 15:21 |
| jeblair | sdague: i think we can agree on ux :) aesthetics is harder to do by committee | 15:22 |
| mordred | jeblair: I unstuck jenkins.o.o for now by hand. fixing it for real is very high on mylist, but I think may take a few iterations - and involve an apt repo | 15:22 |
| jeblair | mordred: ok, thx | 15:22 |
| jeblair | sdague: don't worry about the screenshots; i'm hacking on your change locally | 15:23 |
| sdague | http://img33.imageshack.us/img33/181/94lh.png | 15:23 |
| mordred | jeblair: I think that we may want an apt/yum repo for our machiens that has the latest pip. on further discussion with dstufft, we may want to never install things globally with pip and instead always use either virtualenvs or packages | 15:23 |
| sdague | actually, it's not as bad as I remember, it does end up heavy though | 15:23 |
| mordred | jeblair: but I'm going to play with a bunch of combinations of things and come back with some better data/thoughts | 15:24 |
| sdague | also, my patch didn't actually fix this for the other browsers, so whatever color we land on, I need to fix that | 15:24 |
| sdague | jeblair: #B1D5E3; is an interesting blue that splits the difference, and I think stays in the same hue range better | 15:30 |
| jeblair | sdague: ok, i'll look at it in a sec. i'm also consulting with a graphic designer. | 15:30 |
| sdague | :) | 15:30 |
| sdague | well, here's a fixup with that color. Going to run away for a bit. Feel free to do color tweaking beyond that. | 15:32 |
| openstackgerrit | Sean Dague proposed a change to openstack-infra/config: unified styling on progress bars https://review.openstack.org/41299 | 15:32 |
| *** sgviking has joined #openstack-infra | 15:33 | |
| *** sarob has quit IRC | 15:34 | |
| *** dina_belova has joined #openstack-infra | 15:34 | |
| pleia2 | mordred, fungi - thanks for catching the ssl thing (I did all my testing with self-signed certs, so clearly didn't catch it) I'll try to have a look at the cgit stuff on the flight tonight (today have plans with philly folks before heading home) | 15:39 |
| jeblair | sdague: your color looks good. here's a darker one (based on the link color, but lightened to visually match the density): #6b81a2 if we want something with more contrast. i could go either way, but i think i have a slight preference for the darker one (easier to read at a glance) | 15:42 |
| jeblair | sdague: we could consider a really light zebra stripe row background to help with the scanning. | 15:47 |
| jeblair | sdague: here's the dark color: http://imgur.com/inTi1E3 | 15:49 |
| jeblair | ok. breakfast. | 15:49 |
| mordred | jeblair: I like the dark color | 15:49 |
| mordred | jeblair: I also lke the idea of a zebra stripe... although I do agree it should probably be really light | 15:50 |
| sdague | yeh, I actually strayed away from the dark color intentionally, because on chrome the default is this dark color and it just overwhelmed everything | 15:50 |
| sdague | but, we could roll that way if people want | 15:50 |
| sdague | jeblair: I was actually thinking, the thing you most care about is fails, maybe background color the fail lines | 15:50 |
| sdague | the dark is a nice blue though | 15:51 |
| openstackgerrit | Sean Dague proposed a change to openstack-infra/config: unified styling on progress bars https://review.openstack.org/41299 | 15:52 |
| sdague | ok, that's the change with th 6b blue | 15:52 |
| mordred | sdague: hrm. yeah. background color on fail lines could be nice | 15:54 |
| sdague | yeh, though I think we're going to make it extra noise if we futz with it too much more right now. Get used to the new look for a bit, then figure out if there is a good adjustment to make to make it better. | 15:55 |
| jeblair | sdague: agreed; i'm in no rush. | 15:56 |
| sdague | jeblair: anyway, I updated with the darker blue, it does work nicely with the link color | 15:56 |
| jeblair | cool! | 15:56 |
| sdague | so any time you want to flush it through, that would be cool :) | 15:56 |
| jeblair | sdague: +2d the series | 15:58 |
| openstackgerrit | Monty Taylor proposed a change to openstack-dev/pbr: Fix issue with command escaping on Windows https://review.openstack.org/41283 | 16:14 |
| *** thomasm has joined #openstack-infra | 16:15 | |
| mordred | jeblair, sdague I'm +2 - but do you want me to hold off on APRV? | 16:17 |
| jeblair | mordred: i don't see a reason to hold | 16:17 |
| jeblair | upgrading gearman plugin on jenkins.o.o | 16:19 |
| *** sarob has joined #openstack-infra | 16:22 | |
| openstackgerrit | A change was merged to openstack-infra/config: float the progress bars and statuses to the right https://review.openstack.org/39097 | 16:24 |
| openstackgerrit | A change was merged to openstack-infra/config: unified styling on progress bars https://review.openstack.org/41299 | 16:24 |
| *** dina_belova has quit IRC | 16:24 | |
| *** mberwanger has joined #openstack-infra | 16:26 | |
| *** dina_belova has joined #openstack-infra | 16:27 | |
| *** sarob has quit IRC | 16:28 | |
| jeblair | i upgraded jenkins to the same version that 01 and 02 are running as well | 16:29 |
| *** sarob has joined #openstack-infra | 16:29 | |
| jeblair | mordred, fungi, clarkb: jenkins, jenkins01, jenkins02 are all now running the same version of jenkins and gearman-plugin | 16:29 |
| jeblair | (as is jenkins-dev) | 16:30 |
| mordred | jeblair: woot | 16:32 |
| openstackgerrit | Alessandro Pilotti proposed a change to openstack-dev/pbr: Fixes issue with command escaping on Windows https://review.openstack.org/41283 | 16:33 |
| *** sarob has quit IRC | 16:34 | |
| openstackgerrit | A change was merged to openstack-infra/config: Make subunit2html.py py3k and python2 compatible. https://review.openstack.org/40743 | 16:37 |
| openstackgerrit | A change was merged to openstack-infra/config: Exec pip freeze in gate-requirements-install https://review.openstack.org/38746 | 16:38 |
| openstackgerrit | A change was merged to openstack-infra/config: Add stable/grizzly branch to gerribot for puppet https://review.openstack.org/41002 | 16:40 |
| openstackgerrit | A change was merged to openstack-infra/config: Don't post merges in openstack-dev and openstack-nova https://review.openstack.org/40685 | 16:41 |
| *** Ryan_Lane has quit IRC | 16:44 | |
| openstackgerrit | A change was merged to openstack-infra/config: Add reload support to Asterisk https://review.openstack.org/38665 | 16:44 |
| openstackgerrit | A change was merged to openstack-infra/config: Invert modules.conf logic https://review.openstack.org/38801 | 16:45 |
| *** openstackgerrit has quit IRC | 16:48 | |
| *** openstackgerrit has joined #openstack-infra | 16:49 | |
| openstackgerrit | A change was merged to openstack-dev/pbr: Fix pep8 error https://review.openstack.org/40869 | 16:51 |
| *** dina_belova has quit IRC | 16:52 | |
| openstackgerrit | A change was merged to openstack-dev/pbr: Add option to run testr serially https://review.openstack.org/39811 | 16:54 |
| *** mberwanger has quit IRC | 16:55 | |
| *** jjmb has joined #openstack-infra | 16:56 | |
| *** Ryan_Lane has joined #openstack-infra | 17:15 | |
| *** jjmb has quit IRC | 17:23 | |
| *** Ryan_Lane has quit IRC | 17:24 | |
| *** dina_belova has joined #openstack-infra | 17:34 | |
| *** sarob has joined #openstack-infra | 17:39 | |
| *** sarob has quit IRC | 17:44 | |
| openstackgerrit | Alessandro Pilotti proposed a change to openstack-dev/pbr: Fixes issue with command escaping on Windows https://review.openstack.org/41283 | 17:47 |
| *** SergeyLukjanov has joined #openstack-infra | 17:48 | |
| *** sarob has joined #openstack-infra | 17:57 | |
| *** sarob_ has joined #openstack-infra | 18:01 | |
| *** jjmb has joined #openstack-infra | 18:01 | |
| *** sarob has quit IRC | 18:01 | |
| *** sdake_ has joined #openstack-infra | 18:05 | |
| *** jjmb has quit IRC | 18:22 | |
| *** mriedem has joined #openstack-infra | 18:27 | |
| *** sarob_ has quit IRC | 18:31 | |
| *** sarob has joined #openstack-infra | 18:38 | |
| *** sarob has quit IRC | 18:47 | |
| *** weshay has joined #openstack-infra | 18:51 | |
| *** SergeyLukjanov has quit IRC | 18:59 | |
| *** mjfork has quit IRC | 19:11 | |
| *** sarob has joined #openstack-infra | 19:18 | |
| *** sarob has quit IRC | 19:23 | |
| *** Protux has quit IRC | 19:28 | |
| *** thomasm has quit IRC | 19:40 | |
| openstackgerrit | A change was merged to openstack-infra/jenkins-job-builder: Use a more consistent style for booleans values. https://review.openstack.org/41214 | 19:49 |
| *** dina_belova has quit IRC | 19:52 | |
| *** adalbas has quit IRC | 19:55 | |
| *** adalbas has joined #openstack-infra | 20:07 | |
| *** sarob has joined #openstack-infra | 20:19 | |
| *** weshay has quit IRC | 20:20 | |
| *** sarob has quit IRC | 20:24 | |
| Alex_Gaynor | So there seem to be a couple of zuul sets that got stuck: they've been queued for a while, and aren't running even though there's devstack works available | 20:34 |
| openstackgerrit | Alessandro Pilotti proposed a change to openstack-dev/pbr: Fixes issue with command escaping on Windows https://review.openstack.org/41283 | 20:40 |
| clarkb | jeblair: ^ is that related to the new plugin? | 20:50 |
| *** dina_belova has joined #openstack-infra | 20:53 | |
| *** mriedem has quit IRC | 20:54 | |
| *** mriedem has joined #openstack-infra | 20:55 | |
| *** dina_belova has quit IRC | 20:57 | |
| *** mriedem has quit IRC | 21:00 | |
| *** dina_belova has joined #openstack-infra | 21:03 | |
| *** dina_belova has quit IRC | 21:07 | |
| *** sarob has joined #openstack-infra | 21:19 | |
| *** sarob has quit IRC | 21:24 | |
| *** koobs has joined #openstack-infra | 21:29 | |
| *** dina_belova has joined #openstack-infra | 22:03 | |
| *** dina_belova has quit IRC | 22:08 | |
| *** erfanian has joined #openstack-infra | 22:19 | |
| *** sarob has joined #openstack-infra | 22:20 | |
| *** sarob has quit IRC | 22:25 | |
| *** Protux has joined #openstack-infra | 22:33 | |
| *** Protux has quit IRC | 22:40 | |
| *** boris-42 has quit IRC | 22:41 | |
| *** Protux has joined #openstack-infra | 22:44 | |
| *** Protux has quit IRC | 22:44 | |
| *** Protux has joined #openstack-infra | 22:45 | |
| *** jhesketh has joined #openstack-infra | 22:48 | |
| *** erfanian has quit IRC | 22:49 | |
| *** dina_belova has joined #openstack-infra | 23:04 | |
| *** dina_belova has quit IRC | 23:09 | |
| *** erfanian has joined #openstack-infra | 23:10 | |
| *** mriedem has joined #openstack-infra | 23:16 | |
| *** Ryan_Lane has joined #openstack-infra | 23:19 | |
| *** sarob has joined #openstack-infra | 23:20 | |
| *** jhesketh has quit IRC | 23:23 | |
| *** jhesketh has joined #openstack-infra | 23:23 | |
| *** sarob has quit IRC | 23:25 | |
| *** dnavale has joined #openstack-infra | 23:28 | |
| *** Ryan_Lane has quit IRC | 23:29 | |
| *** Ryan_Lane has joined #openstack-infra | 23:31 | |
| *** fifieldt has joined #openstack-infra | 23:34 | |
| *** dkliban has joined #openstack-infra | 23:51 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
