| openstackgerrit | Ian Wienand proposed openstack/devstack master: [dnm] testing suspected break of setuptools 50 https://review.opendev.org/748934 | 00:38 |
|---|---|---|
| *** cliffparsons has joined #openstack-qa | 00:49 | |
| *** cliffparsons has left #openstack-qa | 00:50 | |
| *** Liang__ has joined #openstack-qa | 01:16 | |
| openstackgerrit | Ian Wienand proposed openstack/devstack master: [wip] Projects in /usr/bin with setuptools 50.0.0 https://review.opendev.org/748937 | 01:19 |
| *** ramishra has joined #openstack-qa | 02:30 | |
| openstackgerrit | Ian Wienand proposed openstack/devstack master: [dnm] Use packaged pip https://review.opendev.org/748942 | 02:48 |
| *** rcernin has quit IRC | 02:58 | |
| *** rcernin has joined #openstack-qa | 03:19 | |
| openstackgerrit | Ian Wienand proposed openstack/devstack master: [dnm] testing suspected break of setuptools 50 https://review.opendev.org/748934 | 03:30 |
| *** rfolco has joined #openstack-qa | 03:31 | |
| *** rcernin has quit IRC | 04:12 | |
| *** rcernin has joined #openstack-qa | 04:14 | |
| *** Liang__ has quit IRC | 04:20 | |
| openstackgerrit | Ian Wienand proposed openstack/devstack master: [dnm] testing suspected break of setuptools 50 https://review.opendev.org/748934 | 04:21 |
| *** evrardjp has quit IRC | 04:33 | |
| *** evrardjp has joined #openstack-qa | 04:33 | |
| *** vishalmanchanda has joined #openstack-qa | 04:35 | |
| *** stephenfin has quit IRC | 05:35 | |
| *** stephenfin has joined #openstack-qa | 05:42 | |
| *** stephenfin has quit IRC | 05:47 | |
| *** rcernin has quit IRC | 05:48 | |
| *** stephenfin has joined #openstack-qa | 05:56 | |
| *** stephenfin has quit IRC | 06:01 | |
| *** stephenfin has joined #openstack-qa | 06:02 | |
| *** rcernin has joined #openstack-qa | 06:03 | |
| *** stephenfin has quit IRC | 06:09 | |
| *** stephenfin has joined #openstack-qa | 06:16 | |
| *** lpiwowar has joined #openstack-qa | 06:17 | |
| openstackgerrit | Ian Wienand proposed openstack/devstack master: [wip] Use SETUPTOOLS_USE_DISTUTILS=stdlib for global pip installs https://review.opendev.org/748957 | 06:23 |
| *** stephenfin has quit IRC | 06:27 | |
| *** stephenfin has joined #openstack-qa | 06:29 | |
| *** gcheresh has joined #openstack-qa | 06:39 | |
| *** stephenfin has quit IRC | 06:50 | |
| *** lpiwowar has quit IRC | 06:52 | |
| *** lpiwowar has joined #openstack-qa | 06:53 | |
| *** stephenfin has joined #openstack-qa | 06:57 | |
| *** ralonsoh has joined #openstack-qa | 07:01 | |
| *** stephenfin has quit IRC | 07:04 | |
| *** gcheresh has quit IRC | 07:10 | |
| *** ccamposr__ has joined #openstack-qa | 07:17 | |
| *** trident has quit IRC | 07:18 | |
| *** ccamposr has quit IRC | 07:20 | |
| frickler | ianw: https://review.opendev.org/#/c/748942 does seem to work fine on focal and fedora, so despite some workarounds being needed for older stuff, I'm thinking that this may be the right way forward | 07:22 |
| *** lpiwowar has quit IRC | 07:24 | |
| *** lpiwowar has joined #openstack-qa | 07:24 | |
| ianw | frickler: yeah, i guess the problem has always been that the distro pip somehow isn't sufficient and we need to go back to upstream | 07:25 |
| ianw | like for example bionic pip not pulling down manylinux hweels | 07:25 |
| frickler | ianw: yeah, but if it works for a year or so, we could then try to bump distros into supporting new stuff if needed | 07:27 |
| ianw | frickler: i guess the concern is when the focal version then breaks, and we're back to the same discussion about how do we pull in upstream again | 07:28 |
| ianw | and yeah, i don't know how to fix the immediate issues with bionic version either | 07:29 |
| frickler | switch to focal for master, cap setuptools for stable branches does sound reasonable to me. but I don't have too much time this week to iron things out, either, so maybe wait and see what others come up with | 07:31 |
| ianw | frickler: yeah, i imagine as the world wakes up the issues will start to come in upstream ... | 07:32 |
| frickler | another option to consider might be dropping the "-e" installs, which seem to be the key trigger in this iiuc. would make life for developers a bit more complicated, but might be worth it overall | 07:39 |
| *** tosky has joined #openstack-qa | 07:39 | |
| ianw | hrrrm, i'm not familiar with that side of it. is that because the .egg or .lnk or whatever it is goes in site-packages? | 07:40 |
| ianw | yeah, it's making /usr/lib/python3.6/site-packages/keystone.egg-link | 07:42 |
| frickler | I'm going to test locally, just read that in some previous discussion. I think clarkb mentioned it for th 48.0 issue? | 07:43 |
| ianw | ... interesting, maybe you're right? | 07:44 |
| *** gcheresh has joined #openstack-qa | 07:44 | |
| ianw | i think it does install to /usr/local/lib/python3.6/dist-packages/keystone without "-e" | 07:45 |
| ianw | maybe it's only "-e" and the .egg-link file getting put in setuptools | 07:45 |
| ianw | frickler : root@158.69.73.143 is one on hold with setuptools 50 | 07:46 |
| *** brinzhang has joined #openstack-qa | 07:49 | |
| *** rcernin has quit IRC | 07:51 | |
| *** tosky has quit IRC | 08:00 | |
| *** ramishra has quit IRC | 08:00 | |
| *** lucasagomes has joined #openstack-qa | 08:01 | |
| *** irclogbot_1 has quit IRC | 08:02 | |
| *** ralonsoh_ has joined #openstack-qa | 08:02 | |
| *** ralonsoh has quit IRC | 08:03 | |
| *** ralonsoh_ has quit IRC | 08:03 | |
| *** tosky has joined #openstack-qa | 08:05 | |
| *** ramishra has joined #openstack-qa | 08:05 | |
| *** irclogbot_0 has joined #openstack-qa | 08:08 | |
| *** lpiwowar has quit IRC | 08:13 | |
| *** lpiwowar has joined #openstack-qa | 08:13 | |
| *** TomStappaerts has joined #openstack-qa | 08:16 | |
| *** ricolin has quit IRC | 08:16 | |
| openstackgerrit | Tom Stappaerts proposed openstack/devstack stable/train: Fix uwsgi install from source https://review.opendev.org/747987 | 08:24 |
| *** dulek has joined #openstack-qa | 08:33 | |
| *** gcheresh has quit IRC | 08:41 | |
| *** gcheresh has joined #openstack-qa | 08:53 | |
| *** trident has joined #openstack-qa | 09:08 | |
| openstackgerrit | Jens Harbott (frickler) proposed openstack/devstack master: DNM: Do not make editable pip installs https://review.opendev.org/749007 | 09:10 |
| *** dtantsur|afk is now known as dtantsur | 09:11 | |
| frickler | ianw: ^^ this worked fine for me locally until it discovered some weird issue in designate-dashboard, lets see how far it gets us in gate | 09:11 |
| slaweq | frickler: ianw: hi, do You know if error like https://zuul.opendev.org/t/openstack/build/9c4f759d06f84887832a9cb18b92498a/log/job-output.txt#11036 is already reported somewhere? | 09:12 |
| slaweq | I see that almost all neutron jobs are failing due to that now | 09:13 |
| ianw | slaweq: yes, see http://lists.openstack.org/pipermail/openstack-discuss/2020-August/016905.html | 09:13 |
| frickler | slaweq: without looking at your job: likely all devstack jobs are broken currently, yes, see ml about setuptools new release | 09:13 |
| slaweq | frickler: ianw thx a lot for confirmation | 09:14 |
| * frickler should likely send a notice about that | 09:14 | |
| ianw | we have a number of options; but also i see more issues coming in on https://github.com/pypa/setuptools/issues/ too | 09:14 |
| *** brinzhang has quit IRC | 09:16 | |
| -openstackstatus- NOTICE: due to a new release of setuptools (50.0.0), a lot of jobs are currently broken, please do not recheck blindly. see http://lists.openstack.org/pipermail/openstack-discuss/2020-August/016905.html | 09:17 | |
| *** brinzhang has joined #openstack-qa | 09:19 | |
| frickler | maybe the final result will be "use venvs for everything you can't get from distro pkgs" | 09:20 |
| ianw | yes, i mean that's probably what we should do | 09:22 |
| ianw | just that's practically never got very far :) | 09:23 |
| *** brinzhang_ has joined #openstack-qa | 09:37 | |
| *** brinzhang has quit IRC | 09:40 | |
| *** ralonsoh has joined #openstack-qa | 09:43 | |
| *** brinzhang_ has quit IRC | 09:52 | |
| *** ramishra has quit IRC | 10:09 | |
| *** ralonsoh has quit IRC | 10:10 | |
| *** ralonsoh has joined #openstack-qa | 10:10 | |
| *** tosky has quit IRC | 10:13 | |
| *** tosky has joined #openstack-qa | 10:13 | |
| *** gfidente has joined #openstack-qa | 10:19 | |
| *** Luzi has joined #openstack-qa | 10:35 | |
| openstackgerrit | Soniya Murlidhar Vyas proposed openstack/tempest master: Need to have stable implementation of nova_volume_detach() and add missing docstrings https://review.opendev.org/749025 | 10:42 |
| *** brinzhang has joined #openstack-qa | 10:43 | |
| openstackgerrit | Tom Stappaerts proposed openstack/devstack stable/train: Fix uwsgi install from source https://review.opendev.org/747987 | 10:45 |
| *** lpiwowar_ has joined #openstack-qa | 10:45 | |
| *** vhari has joined #openstack-qa | 10:45 | |
| *** lpiwowar has quit IRC | 10:47 | |
| *** ramishra has joined #openstack-qa | 10:51 | |
| *** whoami-rajat__ has joined #openstack-qa | 11:12 | |
| *** elod has quit IRC | 11:27 | |
| whoami-rajat__ | hi #openstack-qa , I'm seeing this error in the cinder gate during devstack setup, i couldn't find any direct change that would affect it, can you please take a look http://paste.openstack.org/show/797292/ | 11:30 |
| *** elod has joined #openstack-qa | 11:52 | |
| *** raildo has joined #openstack-qa | 12:12 | |
| frickler | yoctozepto: gmann: what to do think about https://review.opendev.org/749007 vs. capping setuptools in reqs? | 12:17 |
| *** lbragstad_ has joined #openstack-qa | 12:32 | |
| *** redrobot has joined #openstack-qa | 12:34 | |
| openstackgerrit | Lukas Piwowarski proposed openstack/tempest master: Create documentation for tests https://review.opendev.org/735619 | 12:34 |
| frickler | or follow https://review.opendev.org/748957 | 12:36 |
| *** lbragstad_ has quit IRC | 12:39 | |
| *** rh-jelabarre has joined #openstack-qa | 12:39 | |
| *** paras333 has joined #openstack-qa | 12:49 | |
| *** raildo has quit IRC | 12:51 | |
| openstackgerrit | Tom Stappaerts proposed openstack/devstack stable/train: Fix uwsgi install from source https://review.opendev.org/747987 | 12:53 |
| *** lbragstad has joined #openstack-qa | 12:55 | |
| *** raildo has joined #openstack-qa | 13:00 | |
| openstackgerrit | Jens Harbott (frickler) proposed openstack/devstack master: Use SETUPTOOLS_USE_DISTUTILS=stdlib for global pip installs https://review.opendev.org/748957 | 13:04 |
| frickler | yoctozepto: gmann: ianw: clarkb: I guess I'd go with this option ^^ | 13:05 |
| openstackgerrit | Jens Harbott (frickler) proposed openstack/devstack stable/ussuri: Use SETUPTOOLS_USE_DISTUTILS=stdlib for global pip installs https://review.opendev.org/749050 | 13:12 |
| *** artom has joined #openstack-qa | 13:17 | |
| frickler | ^^ the backport conflicted because we still have grenade jobs running as non-voting in gate for ussuri, maybe we remember to clean that up this time ;) | 13:44 |
| *** Luzi has quit IRC | 13:51 | |
| *** nweinber has joined #openstack-qa | 13:56 | |
| *** Liang__ has joined #openstack-qa | 13:57 | |
| *** brinzhang has quit IRC | 14:05 | |
| openstackgerrit | Jens Harbott (frickler) proposed openstack/devstack stable/train: Use SETUPTOOLS_USE_DISTUTILS=stdlib for global pip installs https://review.opendev.org/749063 | 14:09 |
| abhishekk | gmann, glance_store changes related to cinder store are merged now and now we can review this patch https://review.opendev.org/743800 | 14:10 |
| abhishekk | this is the glance patch, which uses above devstack patch https://review.opendev.org/748039 | 14:10 |
| *** Liang__ has quit IRC | 14:31 | |
| *** openstackgerrit has quit IRC | 14:37 | |
| *** TViernion has quit IRC | 14:37 | |
| *** TViernion has joined #openstack-qa | 14:41 | |
| *** gfidente has quit IRC | 14:55 | |
| yoctozepto | frickler: you have my +2, sir :-) | 14:56 |
| yoctozepto | gmann, frickler: btw, for some reason only bionic is affected? I see masakari on focal is passing: https://review.opendev.org/744343 | 14:56 |
| clarkb | as a side note dropping the -e from pip install would likely work too. I expect many people use the -e functionality in devstack so the env var switch is likely a better choice for devstack. But if you run into this in other contexts there are other options | 14:57 |
| frickler | yoctozepto: interesting, I hadn't noticed that before. seems because with install distro python3-setuptools there and then don't upgrade to latest | 15:00 |
| *** lpiwowar_ has quit IRC | 15:00 | |
| *** gfidente has joined #openstack-qa | 15:03 | |
| *** paras333 has quit IRC | 15:04 | |
| yoctozepto | clarkb: env var sounds like fixing the root problem (or at least a bit closer to it), dropping -e like a workaround | 15:04 |
| clarkb | yoctozepto: dropping the -e is not a workaround, its a proper fix | 15:04 |
| yoctozepto | frickler: possibly, or no conflict happens? | 15:04 |
| clarkb | the behavior only happens with -e because distutils is only used with -e | 15:04 |
| clarkb | otherwise setuptools is used without distutils aiui | 15:05 |
| *** paras333 has joined #openstack-qa | 15:05 | |
| yoctozepto | clarkb: ah, interesting, though as you noted we kinda want -e for folks who are accustomed to it | 15:05 |
| yoctozepto | hmm, google is unhelpful https://setuptools.readthedocs.io/en/latest/distutils-legacy.html | 15:06 |
| clarkb | basically -e exposes a distutils feature | 15:07 |
| yoctozepto | https://www.reddit.com/r/Python/comments/ijvtmr/heads_up_recent_setuptools_release_has_broken/ "Don't run pip as root" :D | 15:07 |
| *** mjturek1 has joined #openstack-qa | 15:07 | |
| clarkb | there is a risk that setuptools will be updated to match the distutils behavior though | 15:12 |
| clarkb | which would be more fun bceause then our distutils override would be mismatched with the non -e behavior :) | 15:12 |
| yoctozepto | frickler: yeah, the setuptools seem not to get touched on focal and that's why it's happy | 15:13 |
| yoctozepto | frickler: if we migrated early, we would be green ;d | 15:13 |
| yoctozepto | frickler, clarkb: maybe we should not touch system setuptools at all when doing sudo installs, seems fair but hard to enforce | 15:14 |
| frickler | yoctozepto: clarkb: in the long run, finishing the work to make everything run in venvs may be the best option, like dtantsur mentioned on the ml | 15:15 |
| yoctozepto | or just go dtantsur's idea and venv it all | 15:15 |
| frickler | yoctozepto: ^5 ;) | 15:15 |
| yoctozepto | frickler: you read my mind | 15:15 |
| clarkb | yup, I reached that conclusion back when pip 10 happened | 15:15 |
| clarkb | but there are a lot of difficulties :/ | 15:15 |
| yoctozepto | well, we can just use kolla ;d | 15:16 |
| yoctozepto | Sean Mooney likes to dev with kolla so it might not be that bad of an idea | 15:16 |
| yoctozepto | but not for today obviously | 15:16 |
| clarkb | it come sup frequently. My major concerns aboutdoing that are the very large overheads compared to regular devstack which lead to even longer setuptimes aiui | 15:16 |
| clarkb | also we'd lose co installability checking | 15:17 |
| clarkb | devstack is really slow itself, the bar isn't set high but it seems the other tools don't really do better | 15:17 |
| yoctozepto | clarkb: true that, as they care even more about production features | 15:18 |
| yoctozepto | rather than raw speed for development | 15:18 |
| frickler | maybe some day nobody is going to need co-installability anymore. think openstack-operator | 15:18 |
| yoctozepto | frickler: kolla does not need it and hence why it does not enforce/test it ;d | 15:18 |
| yoctozepto | but anyhow | 15:18 |
| yoctozepto | what are we doing right here and right now? | 15:19 |
| clarkb | it sounds like setting that env var? | 15:19 |
| yoctozepto | sad gmann not around, I'd just merge that gate greenifier right away | 15:19 |
| clarkb | we can always revert and do some other change if necessary | 15:19 |
| frickler | https://review.opendev.org/748957 just passed, clarkb do you want to review otherwise I'd self-approve | 15:20 |
| clarkb | for the grenade question getting into the habit of forward porting fixes would be good | 15:20 |
| clarkb | it simplifies cleanup | 15:20 |
| clarkb | do we want to push up a bunch of stable branch changes to do that or deal with the cleanup of marking the jobs voting again once it is working? | 15:20 |
| frickler | clarkb: the stable branches haven't been cleaned up yet anyway from the last event | 15:21 |
| frickler | I'll push changes for that once the current set gets in | 15:21 |
| frickler | https://review.opendev.org/749050 seems ready, too | 15:21 |
| clarkb | frickler: ok I'll approve the env var chnage then | 15:22 |
| yoctozepto | thanks | 15:22 |
| dtantsur | I don't want to get you scared, but we seem to see this breakage even with venv.. | 15:22 |
| dtantsur | (not on devstack, on bifrost) | 15:22 |
| yoctozepto | dtantsur: :O | 15:22 |
| yoctozepto | well, that's counter-logic | 15:22 |
| dtantsur | it may be a separate fallout from the same setuptools release | 15:22 |
| frickler | dtantsur: same break or different? do you have a pointer? | 15:22 |
| clarkb | I'm not sure how this would happen wit ha virtualenv | 15:23 |
| frickler | we have seen some other failure with venv on py35 in #opendev | 15:23 |
| clarkb | since the virtualenv bin dir is self contained (and independent of the distro) | 15:23 |
| clarkb | it may be there is a different regression | 15:23 |
| dtantsur | frickler: https://zuul.opendev.org/t/openstack/build/86049990893a40f1a3732485d2226623/log/job-output.txt#9553 happens only on focal | 15:23 |
| dtantsur | ModuleNotFoundError: No module named 'setuptools._distutils'\n\n" | 15:24 |
| yoctozepto | hah, only on focal you say... | 15:24 |
| yoctozepto | ;D | 15:24 |
| dtantsur | I *think* we have to use --system-site-packages | 15:24 |
| dtantsur | but I'm probably distracting you :) | 15:26 |
| yoctozepto | that would be a bad decision | 15:26 |
| yoctozepto | nah, it's the right thing to discuss in here | 15:26 |
| dtantsur | yoctozepto: there is no way around it if you want to use ansible and install packages | 15:26 |
| dtantsur | yeah | 15:26 |
| yoctozepto | dtantsur: you can link up the selinux if that's what you are about | 15:26 |
| dtantsur | yoctozepto: I'm about python-apt/python-dnf not distributed via pypi | 15:27 |
| yoctozepto | dtantsur: ah, those, I *think* the same applies, but it's a bit tricker afair selinux is more self-contained | 15:28 |
| dtantsur | anyway, please move on :) | 15:28 |
| * frickler moves on and afks, bbl | 15:28 | |
| * yoctozepto moves on to masakari again | 15:29 | |
| gmann | yoctozepto: frickler sorry, just finished my internal meetings | 15:30 |
| *** ricolin has joined #openstack-qa | 15:31 | |
| yoctozepto | gmann: no problem, we are just merging one of the fixes and wanted your feedback too | 15:34 |
| gmann | +1, read the log. | 15:41 |
| *** ricolin has quit IRC | 16:00 | |
| *** lucasagomes has quit IRC | 16:02 | |
| *** ricolin has joined #openstack-qa | 16:08 | |
| *** artom has quit IRC | 16:10 | |
| *** openstackgerrit has joined #openstack-qa | 16:19 | |
| openstackgerrit | Oleksandr Kozachenko proposed openstack/devstack master: Remove legacy special cases in is_service_enabled https://review.opendev.org/749082 | 16:19 |
| openstackgerrit | James Parker proposed openstack/whitebox-tempest-plugin master: Update exception message for nova service https://review.opendev.org/749085 | 16:27 |
| *** gfidente has quit IRC | 16:49 | |
| *** ralonsoh has quit IRC | 16:53 | |
| *** dtantsur is now known as dtantsur|afk | 17:13 | |
| *** tosky has quit IRC | 17:14 | |
| *** ccamposr has joined #openstack-qa | 17:17 | |
| *** ccamposr__ has quit IRC | 17:20 | |
| *** artom has joined #openstack-qa | 17:49 | |
| artom | jparker_, so, the failure I'm seeing is "ModuleNotFoundError: No module named 'neutron'" | 17:50 |
| artom | Which looks like a devstack failure... | 17:50 |
| * jparker_ looking | 17:50 | |
| jparker_ | artom: oh sorry I was looking at the compute logs not the controller | 17:51 |
| clarkb | its probably more fallout from the setuptools update (and maybe the fix? though that hasn't landed yet I don't think) | 17:52 |
| artom | clarkb, fix doesn't look to have landed yet... | 17:54 |
| *** rfolco has quit IRC | 17:54 | |
| artom | I guess we need to wait for https://review.opendev.org/#/c/748957/2 to get in? | 17:55 |
| gmann | clarkb: frickler ianw we need to disable grenade job in template too to merge devstack fix - https://review.opendev.org/#/c/749095/1 | 18:01 |
| gmann | artom: yes | 18:01 |
| clarkb | or do the forward porting | 18:01 |
| clarkb | I personally prefer forward porting to deal with grenade | 18:01 |
| clarkb | but I need to pop out now for a bit so will defer to ohters | 18:01 |
| gmann | ok, i am ok with that. train one should pass | 18:02 |
| gmann | +2 on train one too - https://review.opendev.org/#/c/749063/1 | 18:03 |
| gmann | yoctozepto: ^^ | 18:03 |
| clarkb | oh do we have backports already? | 18:03 |
| clarkb | I can review those really quickly before popping out | 18:03 |
| gmann | clarkb: yeah | 18:03 |
| *** rfolco has joined #openstack-qa | 18:08 | |
| *** rfolco has quit IRC | 18:14 | |
| openstackgerrit | Doug Schveninger proposed openstack/patrole master: Clean up tox.ini and document zuul.yaml settings https://review.opendev.org/747806 | 18:23 |
| openstackgerrit | Hemachandra Reddy proposed openstack/tempest master: Wait until dependent resource is deleted https://review.opendev.org/748765 | 18:28 |
| frickler | gmann: ah, now I know why grenade was made n-v in ussuri gate instead of commented out :-/ I think we can approve ussuri now anyway, because it will be on top of train in the integrated gate queue | 18:32 |
| gmann | frickler: yeah | 18:37 |
| gmann | frickler: done | 18:38 |
| openstackgerrit | James Parker proposed openstack/whitebox-tempest-plugin master: Added instance collocation migration test https://review.opendev.org/704420 | 19:48 |
| openstackgerrit | Merged openstack/devstack stable/train: Use SETUPTOOLS_USE_DISTUTILS=stdlib for global pip installs https://review.opendev.org/749063 | 19:51 |
| openstackgerrit | Merged openstack/devstack stable/ussuri: Use SETUPTOOLS_USE_DISTUTILS=stdlib for global pip installs https://review.opendev.org/749050 | 20:01 |
| *** tosky has joined #openstack-qa | 20:15 | |
| *** nweinber has quit IRC | 20:17 | |
| *** paras333 has quit IRC | 20:31 | |
| *** ccamposr__ has joined #openstack-qa | 20:51 | |
| *** paras333 has joined #openstack-qa | 20:51 | |
| *** ccamposr has quit IRC | 20:52 | |
| *** paras333 has quit IRC | 20:56 | |
| *** gcheresh has quit IRC | 21:00 | |
| *** ccamposr__ has quit IRC | 21:01 | |
| clarkb | dtantsur|afk: fyi I think your issue with setuptools 50 may be https://github.com/pypa/setuptools/issues/2352 | 21:09 |
| clarkb | or related anyway I think you're on a newe rpython but the failure is similar | 21:09 |
| ianw | so we merged both the env var fix and the requirements change? | 21:10 |
| clarkb | there was a requirements change? | 21:11 |
| clarkb | (I think requirements ar eusually too late) | 21:11 |
| gmann | i did not know either | 21:11 |
| ianw | yeah -- it's sort of a corner case because we re-install setuptools in install_pip.sh | 21:12 |
| clarkb | ah | 21:12 |
| ianw | https://review.opendev.org/#/c/748943/ | 21:13 |
| ianw | and yeah, as later discussion found, that changelog isn't i guess 100% correct. i does install with "-e" into site-packages, but not without "-e" | 21:13 |
| ianw | just to be super confusing | 21:13 |
| clarkb | ya its definitely -e specific | 21:14 |
| *** vishalmanchanda has quit IRC | 21:15 | |
| ianw | i do hope that change merging hasn't hidden any defects that SETUPTOOLS_USE_DISTUTILS might have | 21:15 |
| ianw | are we sure that hasn't run against setuptools <50? ... checking | 21:16 |
| ianw | i guess on stable branches i could not have | 21:16 |
| clarkb | ya unless we backported to stable too | 21:19 |
| ianw | no it wasn't, but i still have some concerns ... | 21:20 |
| gmann | this is with 50 on master (before req was merged) https://zuul.opendev.org/t/openstack/build/c8b572906e3a4ad0aae36ecbbc95bc7d/log/job-output.txt#3542 | 21:21 |
| ianw | 2020-08-31 15:13:52.886 | Successfully installed pip-9.0.3 setuptools-44.1.1 wheel-0.35.1 | 21:22 |
| ianw | https://storage.gra.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_09c/749063/1/check/devstack/09c6df3/controller/logs/devstacklog.txt | 21:22 |
| ianw | that's the train job | 21:22 |
| clarkb | ianw: was that on python2? | 21:22 |
| clarkb | I think newer setuptools is python3 only? | 21:22 |
| gmann | ussuri one is also with 50 | 21:22 |
| ianw | ok, so train does a cap-pip.txt to <10 | 21:23 |
| ianw | https://storage.bhs.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_33a/749050/1/check/devstack/33ad688/controller/logs/devstacklog.txt | 21:27 |
| ianw | that ussuri | 21:27 |
| ianw | sudo -H LC_ALL=en_US.UTF-8 SETUPTOOLS_USE_DISTUTILS=stdlib http_proxy= https_proxy= no_proxy= PIP_FIND_LINKS= SETUPTOOLS_SYS_PATH_TECHNIQUE=rewrite /usr/local/bin/pip3.6 install -c /opt/stack/requirements/upper-constraints.txt 'setuptools!=24.0.0,!=34.0.0,!=34.0.1,!=34.0.2,!=34.0.3,!=34.1.0,!=34.1.1,!=34.2.0,!=34.3.0,!=34.3.1,!=34.3.2,!=36.2.0;python_version>='\''3.5'\''' 'setuptools!=24.0.0,!=34.0.0,!=34.0.1,!=34.0.2,!=34.0.3,!=34.1.0,!=34.1.1,!=34. | 21:27 |
| ianw | 2.0,!=34.3.0,!=34.3.1,!=34.3.2,!=36.2.0,<45.0.0;python_version<='\''2.7'\''' | 21:27 |
| *** ccamposr has joined #openstack-qa | 21:27 | |
| ianw | so i'd agree that has installed 50 | 21:28 |
| ianw | as gmann said :) | 21:28 |
| ianw | i think then we can probably revert the requirements change | 21:28 |
| clarkb | as long we we base it on the master env var change? | 21:29 |
| ianw | clarkb: so master is the one i'd want to verify is working *without* the requirements change | 21:33 |
| clarkb | ya so pushing up a revert on top of the env var change should confirm that? | 21:34 |
| ianw | https://zuul.opendev.org/t/openstack/build/ecc2f16c7501459fb566c59c670dd753/log/controller/logs/devstacklog.txt#1836 | 21:34 |
| ianw | this was the gate job, which didn't seem to run with that requirements change | 21:35 |
| ianw | pip_install 'setuptools!=24.0.0,!=34.0.0,!=34.0.1,!=34.0.2,!=34.0.3,!=34.1.0,!=34.1.1,!=34.2.0,!=34.3.0,!=34.3.1,!=34.3.2,!=36.2.0,!=48.0.0,!=49.0.0' | 21:35 |
| ianw | i mean WITHOUT that requirements change, sorry | 21:35 |
| ianw | i.e. i think we're good | 21:35 |
| ianw | the requirements change isn't hiding any failures with the SETUPTOOLS_USE_DISTUTILS change | 21:36 |
| gmann | yeah | 21:40 |
| *** raildo has quit IRC | 21:41 | |
| *** ccamposr has quit IRC | 21:49 | |
| *** ccamposr__ has joined #openstack-qa | 21:50 | |
| *** ccamposr__ has quit IRC | 21:50 | |
| *** ccamposr__ has joined #openstack-qa | 21:51 | |
| clarkb | has the requirements change merged? I'm looking at my local git tree and seems like no? | 21:52 |
| gmann | yeah around 4hrs before - https://review.opendev.org/#/c/748943/ | 21:54 |
| ianw | revert proposed @ https://review.opendev.org/749159 | 21:55 |
| clarkb | oh it was in requirements no wonder I'm confused (I thought you meant it was in devstack's get-pip.py instal list but cleraly when you said requirements that is what you meant :P) | 21:56 |
| *** whoami-rajat__ has quit IRC | 22:01 | |
| *** ccamposr__ has quit IRC | 22:06 | |
| *** ccamposr__ has joined #openstack-qa | 22:06 | |
| ianw | yeah we have the cap-pip.txt, but then also the setuptools re-install from requirements | 22:06 |
| ianw | i'm trying to decide if i should look into this more or just see what comes out of upstream; which is currently not much response to the several issues | 22:09 |
| *** slaweq has quit IRC | 22:15 | |
| clarkb | they were responsive when setuptools 48 broke things | 22:16 |
| clarkb | maybe they just need a day | 22:17 |
| *** artom has quit IRC | 22:30 | |
| *** artom has joined #openstack-qa | 22:31 | |
| *** stephenfin has joined #openstack-qa | 22:37 | |
| ianw | so the "-e" thing i think is a red-herring; i think by default without "-e" you're building a wheel and installing that | 22:40 |
| ianw | which seems to take a different path that gets things right | 22:41 |
| clarkb | ianw: it isn't a red herring its directly related aiui. The original bug from setuptools 48 goes into details | 22:41 |
| clarkb | basically -e is an editable install which depends on distutils to do its thing | 22:41 |
| clarkb | if you don't do an editable install then you bypass distutils entirely | 22:42 |
| ianw | right, yeah but i mean "pip install --no-binary :all: ." also installs it incorrectly | 22:42 |
| clarkb | oh interesting, I wouldn't have expected that to go into distutils either but I guess it does | 22:42 |
| *** rcernin has joined #openstack-qa | 22:51 | |
| *** tosky has quit IRC | 23:03 | |
| *** paras333 has joined #openstack-qa | 23:47 | |
| openstackgerrit | Doug Schveninger proposed openstack/patrole master: Fix Extension gates by supporting Skip Execption in test case. https://review.opendev.org/746480 | 23:48 |
| openstackgerrit | Merged openstack/devstack master: Use SETUPTOOLS_USE_DISTUTILS=stdlib for global pip installs https://review.opendev.org/748957 | 23:55 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!