| opendevreview | Julia Kreger proposed openstack/ironic-inspector master: WIP SQLAlchemy 2.0 prep https://review.opendev.org/c/openstack/ironic-inspector/+/860731 | 00:34 |
|---|---|---|
| vanou | good morning ironic | 02:20 |
| opendevreview | OpenStack Proposal Bot proposed openstack/ironic-inspector master: Imported Translations from Zanata https://review.opendev.org/c/openstack/ironic-inspector/+/861023 | 03:16 |
| opendevreview | OpenStack Proposal Bot proposed openstack/ironic master: Imported Translations from Zanata https://review.opendev.org/c/openstack/ironic/+/861027 | 03:25 |
| arne_wiebalck | Good morning, Ironic! | 06:34 |
| jm1 | dtantsur: good morning :) maybe you could have another look at this patch please? 👀 https://review.opendev.org/c/openstack/bifrost/+/859430/ | 07:26 |
| rpittau | good morning ironic! o/ | 07:40 |
| opendevreview | Riccardo Pittau proposed openstack/metalsmith stable/zed: Allow to use Ansible OpenStack Col. 1.x.x with openstacksdk >=0.99.0 https://review.opendev.org/c/openstack/metalsmith/+/861050 | 08:14 |
| rpittau | any core with good heart having a minute to dedicate to fix metalsmith CI in zed reviewing this ^ would be greatly appreciated :) | 08:25 |
| dtantsur | jm1: just back from PTO, will put on my queue | 09:13 |
| jm1 | dtantsur: thank you :) btw you dont want to know which dirty hacks we put into metalsmith while you were afk :D | 09:15 |
| dtantsur | ouch ouch | 09:15 |
| dtantsur | what will win: my curiosity or my desire to keep sanity? | 09:15 |
| * rpittau hides | 09:15 | |
| jm1 | dtantsur: in case curiosity wins, just follow rpittau's link above ;) | 09:16 |
| jm1 | dtantsur: btw we can remove this hack once we backport some of the baremetal patches to aoc's stable branch | 09:17 |
| jm1 | dtantsur: but thats work in progress | 09:17 |
| opendevreview | Merged openstack/ironic-inspector master: Imported Translations from Zanata https://review.opendev.org/c/openstack/ironic-inspector/+/861023 | 10:38 |
| opendevreview | Merged openstack/ironic master: Imported Translations from Zanata https://review.opendev.org/c/openstack/ironic/+/861027 | 10:44 |
| opendevreview | Ebbex proposed openstack/bifrost master: Enable epel repository for more than CentOS https://review.opendev.org/c/openstack/bifrost/+/861082 | 12:43 |
| TheJulia | dtantsur: any opinion if we rm -f ironic_inspector/tests/functional.py ? | 13:43 |
| dtantsur | :D | 13:43 |
| dtantsur | TheJulia: chances are non-zero, it exercises things that are not covered by tempest | 13:43 |
| dtantsur | why? | 13:43 |
| TheJulia | it is being *exceptionally* problematic with sqlalchemy 2.0 and turning essentially requires autocommit to work | 13:43 |
| TheJulia | s/turning// | 13:44 |
| TheJulia | even pointing it at mysql, is... abysmal with sqlalchemy 2.0 :( | 13:44 |
| dtantsur | this is pretty weird, I think it only needs to rebuild the database | 13:45 |
| TheJulia | how many human-hours is it worth to try and keep it, do you think? | 13:45 |
| dtantsur | I'm a bit worried about the situation itself, rather than the specific tests | 13:46 |
| TheJulia | (I wouldn't be asking if I hadn't already lost two days to it) | 13:46 |
| dtantsur | it feels like we're ready to sacrifice anything just to be able to use SQLAlchemy 2 | 13:46 |
| TheJulia | two days of banging my head | 13:46 |
| dtantsur | My question in response: at which point do we start to push back to upgrading to it? | 13:46 |
| TheJulia | dunno, tempest/integration seem to work just fine... although I broke some of the unit tests on my wip patch and I'm having to fix those now since I kept trying to get functional tests working | 13:48 |
| dtantsur | if we remove functional tests, we'll also completely lose coverage for ironic-inspector-client | 13:49 |
| dtantsur | both may be fine since ironic-inspector barely sees any changes any more | 13:49 |
| TheJulia | the functional tests don't use ironic-inspector-client at all | 13:50 |
| dtantsur | TheJulia: there is a version in irnoic-inspector-client that does | 13:50 |
| TheJulia | part of the reason I'm curious how much you'd desire to keep it | 13:50 |
| TheJulia | ... what? | 13:50 |
| dtantsur | https://opendev.org/openstack/python-ironic-inspector-client/src/branch/master/ironic_inspector_client/tests/functional.py | 13:51 |
| dtantsur | it's derived from the inspector's functional tests | 13:51 |
| TheJulia | oh jeeze | 13:51 |
| dtantsur | yes, it's ugly, and yes, I'm to blame :) | 13:51 |
| TheJulia | Is it too early to add liquor to coffee? | 13:51 |
| dtantsur | I'm not going to shed a single tear if these are gone | 13:51 |
| dtantsur | I guess my biggest frustration is that we probably don't understand our own DB layer well enough | 13:52 |
| TheJulia | Well, that makes me feel a little better | 13:52 |
| dtantsur | (you don't understand well enough, everyone else doesn't understand at all :D) | 13:52 |
| TheJulia | (more like... we didn't understand it before hand, and now we're in sqlalchemy locking dungeon) | 13:53 |
| dtantsur | yeah, I don't claim we've ever understood it :) | 13:53 |
| TheJulia | ((and sqlite is monologing on how we will meet our doom....)) | 13:54 |
| * TheJulia is sure a superhero is there someplace | 13:54 | |
| dtantsur | I do hope we don't break sqlite support completely in the end :) | 13:54 |
| TheJulia | ... we likely need to have a scenario job in place someplace, because it is not about our support as much as it is autocommit going away | 13:55 |
| TheJulia | inspector's state machine appears to tickle things sufficently that it is easy to deadlock | 13:56 |
| JayF | dtantsur: We cannot skip on the SQLA2.0 upgrade. Not unless we want to be stuck on old software indefinately. Not upgrading sqlalchemy is a hole that is extremely difficult to get out of, I've worked on multiple projects that tried to freeze it and in the end it was horrible. | 14:29 |
| dtantsur | It's less about staying on 1.0 forever, more about: if the migration is taking so much effort, it's alarming | 14:30 |
| JayF | It's a pretty aggressive API change by SQLA, honestly | 14:32 |
| JayF | and I suspect that is to *force us to understand the DB layer we never understood* to prevent hidden bugs | 14:32 |
| ebbex | dtantsur: https://review.opendev.org/c/openstack/bifrost/+/741964/26/playbooks/roles/bifrost-ironic-install/defaults/required_defaults_Debian_family.yml Do you remember why apache2-utils is installed? | 14:45 |
| ebbex | I'm trying to do some mapping/cleanup of the required_defaults, (ref. https://review.opendev.org/c/openstack/bifrost/+/855805/1//COMMIT_MSG#15) and I don't really see the reason why many of these are listed here. like py3-{dev,pip,setuptools} are all handled earlier on, py3-{iniparse,pymysql} feels like they should be a venv install, not a system package, much like passlib in lower-constraints. | 15:00 |
| dtantsur | ebbex: apache2-utils.. probably for htpasswd? | 15:00 |
| ebbex | I don't see the need for htpasswd though? The ansible htpass module which we use requires passlib. | 15:02 |
| dtantsur | hmm | 15:02 |
| ebbex | if I'm not using dib, the list of required_packages i actually need on deb/alma/suse is [ipmitool, ipxe, isolinux, mariadb-server, dnsmasq and iptables]. | 15:05 |
| ebbex | maybe selinux/policycoreutils on alma, but those can probably be in the venv aswell. | 15:08 |
| TheJulia | JayF: oh, it absolutely is designed to push people to better patterns and to try and understand what is happening under the hood. | 15:10 |
| rpittau | bye everyone, have a great evening! o/ | 16:09 |
| TheJulia | JayF: w/r/t allocations table, your testing revealed that the update migrated the table?! | 16:26 |
| JayF | TheJulia: yes | 16:26 |
| JayF | which should be a noop if it's utf8 -> utf8mb3 | 16:26 |
| TheJulia | just the character encoding? or character encoding and engine? | 16:26 |
| TheJulia | yah | 16:26 |
| JayF | I did not test engine; I already have experienced that behavior (it won't migrate engine) | 16:26 |
| TheJulia | yeah, that is what I was thinking | 16:26 |
| TheJulia | going to add a patch to check the engine and warn upon it, since some mariadb builds out in the wild don't even support myiasm | 16:27 |
| JayF | If it's possible, we should detect + loudly warn if encoding will be migrated from anything but utf8, same for engine | 16:27 |
| * TheJulia looks at her local db | 16:27 | |
| JayF | and make sure upgrade notes indicate that this migration exists, because I wonder if it's *really* noop for very large DBs | 16:27 |
| TheJulia | JayF: writing an explicit upgrade check to do so | 16:27 |
| JayF | even updating the metadata if there is a large amount of rows may not be free (?)... I'm just not sure | 16:28 |
| JayF | and I don't want someone to get surprised | 16:28 |
| opendevreview | Merged openstack/ironic-inspector stable/wallaby: CI: Zuul no longer respects queue https://review.opendev.org/c/openstack/ironic-inspector/+/860163 | 16:34 |
| opendevreview | Merged openstack/ironic bugfix/19.0: Stable only: Correct release note, packaging not required https://review.opendev.org/c/openstack/ironic/+/860740 | 16:34 |
| opendevreview | Merged openstack/ironic bugfix/18.1: Stable only: Correct release note, packaging not required https://review.opendev.org/c/openstack/ironic/+/860742 | 16:34 |
| opendevreview | Merged openstack/ironic stable/xena: Stable only: Correct release note, packaging not required https://review.opendev.org/c/openstack/ironic/+/860741 | 16:34 |
| opendevreview | Merged openstack/ironic stable/yoga: Stable only: Correct release note, packaging not required https://review.opendev.org/c/openstack/ironic/+/860738 | 16:34 |
| opendevreview | Merged openstack/ironic-python-agent-builder stable/wallaby: dhcp-all-interfaces: let NetworkManager doit. https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/860433 | 17:23 |
| JayF | Going to post some things here that are passing tests, and fix CI, and just need landing: | 17:38 |
| JayF | https://review.opendev.org/c/openstack/ironic-inspector/+/860164 | 17:38 |
| JayF | https://review.opendev.org/c/openstack/ironic-inspector/+/860165 | 17:38 |
| JayF | TheJulia: I think some of those rel note updates got merged w/o the acutal code it was describing; e.g. https://review.opendev.org/c/openstack/ironic/+/860142 | 17:40 |
| TheJulia | JayF: if you sign | 17:40 |
| TheJulia | err | 17:40 |
| TheJulia | sigh | 17:40 |
| JayF | it's not a big deal; I'm trying to get those landed all today so I can revamp the release pr | 17:41 |
| JayF | I just noticed it | 17:41 |
| TheJulia | https://review.opendev.org/c/openstack/ironic/+/860144 <-- wut | 17:43 |
| opendevreview | Jay Faulkner proposed openstack/ironic stable/wallaby: Stable only: Factor out addition of packaging lib https://review.opendev.org/c/openstack/ironic/+/860144 | 17:44 |
| TheJulia | the depends-on... | 17:44 |
| JayF | what depends-on <.< >.> | 17:44 |
| JayF | lol | 17:44 |
| TheJulia | lol | 17:44 |
| JayF | it was depending on the ironic CI fixes | 17:44 |
| JayF | but that got landed | 17:44 |
| JayF | and I probably screwed up the syntax | 17:44 |
| JayF | still developing my skills for dealing with over 9000 PRs lol | 17:44 |
| TheJulia | https://review.opendev.org/c/openstack/ironic/+/860144/3/requirements.txt | 17:45 |
| TheJulia | oh, nvmd | 17:45 |
| JayF | yeah, it shows two lines changed b/c I didn't have a newline at the end | 17:45 |
| JayF | which is not ideal but also doesn't matter | 17:45 |
| TheJulia | yup | 17:46 |
| JayF | thanks for approving all those, now just hoping CI will work on 'em lol | 17:46 |
| JayF | https://review.opendev.org/c/openstack/ironic-inspector/+/860170 any ideas on how to get CI on this one happy? I thought the pins I put in tox.ini would be enough but I'm clearly missing something | 17:49 |
| TheJulia | eww | 17:50 |
| JayF | hey, milestone: my open patches list fits on my entire monitor again lol | 17:52 |
| TheJulia | \o. | 17:53 |
| TheJulia | err, \o/ | 17:53 |
| JayF | Oh hey, I should say this here: tomorrow I'm going to start a new thing; Office Hours. Basically live streaming my "normal work" for a couple of hours on youtube, while being around if anyone wants to ask questions or do interactive reviews, whatever. See https://twitter.com/jayofdoom/status/1580246189974654977 for more info | 17:54 |
| TheJulia | well, the most recent failure is definitely more along the lines of tests are not happy | 17:55 |
| JayF | The conflict is caused by: | 17:58 |
| JayF | The user requested automaton>=1.9.0 | 17:58 |
| JayF | The user requested (constraint) automaton===3.0.1 | 17:58 |
| JayF | is what I'm seeing for all those tests | 17:58 |
| JayF | which to me implies some kind of bad requirement being pulled in | 17:58 |
| JayF | hmm | 17:58 |
| JayF | OH | 17:58 |
| JayF | Hmm. | 17:59 |
| JayF | > - openstack-lower-constraints-master-branch-jobs | 17:59 |
| JayF | looks suspicious | 17:59 |
| JayF | I feel like we need to actually pin openstack/requirements somewhere in the zuul jobs yaml? | 17:59 |
| TheJulia | l-c jobs are no more, but it shouldn't be contaminating that one, unless it is the lc job on that | 17:59 |
| JayF | because it looks like that unit test is using requirements.txt from devstack (not from the default location in tox.ini) | 18:00 |
| JayF | which is why my tox.ini change didn't help | 18:00 |
| TheJulia | the fact https://zuul.opendev.org/t/openstack/build/cb901a402b9d4ae4865386263c2f5ca3 is not the case on all jobs is suspect | 18:03 |
| JayF | [3432] /home/zuul/src/opendev.org/openstack/ironic-inspector$ /home/zuul/src/opendev.org/openstack/ironic-inspector/.tox/py36/bin/python -m pip install -c/home/zuul/src/opendev.org/openstack/requirements/upper-constraints.txt -r/home/zuul/src/opendev.org/openstack/ironic-inspector/requirements.txt -r/home/zuul/src/opendev.org/openstack/ironic-inspector/test-requirements.txt | 18:04 |
| JayF | note the URL to upper-constraints being passed | 18:04 |
| JayF | unless we somhow limit the requirements version going in there somewhere else | 18:04 |
| JayF | it's using u-c from master and requirements from bugfix/whateverit is | 18:04 |
| TheJulia | but yeah, definitely using master reqs | 18:04 |
| TheJulia | also the job is using older ubuntu | 18:05 |
| TheJulia | which means the mirror of wheels does not have a prebuilt 3.0.1 wheel | 18:05 |
| TheJulia | weird... | 18:05 |
| TheJulia | there is a way to override it, I'm jut not sure | 18:05 |
| JayF | I'm going to try to now | 18:05 |
| TheJulia | rpittau: ^^^ I seem to remember you've run into something like this before, if we haven't solved it by tomorrow, help would be appreciated | 18:06 |
| JayF | I mean, I'm going to punt on the inspector ones if we have anyone who might have real-knowledge about tackling this | 18:08 |
| JayF | because I suspect there are still lower-hanging-fruit on the changes I have up, and I want to merge in the ones that are close | 18:08 |
| JayF | before trying to figure out the hard cases | 18:08 |
| JayF | https://review.opendev.org/c/openstack/ironic/+/860851 and https://review.opendev.org/c/openstack/ironic/+/860852 are a couple of good changes, I don't wanna land with single +2, this is Julia's changes coming out of the last meeting to reduce testing on t/u/v | 18:20 |
| opendevreview | Merged openstack/ironic-inspector stable/victoria: CI: Various fixes https://review.opendev.org/c/openstack/ironic-inspector/+/860164 | 18:45 |
| opendevreview | Merged openstack/ironic-inspector stable/ussuri: CI: Zuul no longer respects queue https://review.opendev.org/c/openstack/ironic-inspector/+/860165 | 18:45 |
| opendevreview | Merged openstack/ironic bugfix/18.1: Stable only: Factor out addition of packaging lib https://review.opendev.org/c/openstack/ironic/+/860143 | 18:45 |
| opendevreview | Julia Kreger proposed openstack/ironic master: Add upgrade check warning for allocations db https://review.opendev.org/c/openstack/ironic/+/861107 | 20:11 |
| opendevreview | Verification of a change to openstack/ironic stable/xena failed: Stable only: Factor out addition of packaging lib https://review.opendev.org/c/openstack/ironic/+/860142 | 20:38 |
| opendevreview | Julia Kreger proposed openstack/ironic master: Fix allocations default table type https://review.opendev.org/c/openstack/ironic/+/860198 | 20:58 |
| opendevreview | Julia Kreger proposed openstack/ironic master: Phase 1 - SQLAlchemy 2.0 Compatability https://review.opendev.org/c/openstack/ironic/+/856336 | 20:59 |
| opendevreview | Julia Kreger proposed openstack/ironic master: Phase 2 - SQLAlchemy 2.0 Compatability https://review.opendev.org/c/openstack/ironic/+/857516 | 20:59 |
| opendevreview | Julia Kreger proposed openstack/ironic master: Phase 3 - SQLAlchemy 2.0 Compatability https://review.opendev.org/c/openstack/ironic/+/857932 | 20:59 |
| TheJulia | okay, mostly just restacking those^ with the exception of the first | 20:59 |
| JayF | ah, someone needs to kick ironicbaremetal.org to be aware of the zed release | 21:01 |
| JayF | I should probably do that today if nobody else is gonna | 21:01 |
| TheJulia | k | 21:01 |
| * JayF getting some proofs printed of these -> https://home.jvf.cc/~jay/Ironic_mascot_with_qr.svg as die cut stickers (qr code goes to ironicbaremetal.org) | 21:02 | |
| TheJulia | I'm likely going to head out soon after uploading current change states to gerrit | 21:02 |
| JayF | if anyone wants to get on the list to get some, just reach out to me | 21:02 |
| JayF | I'm getting a bunch printed after validating the QR code works on a proof, courtesy of G-Research Open Source Software team | 21:02 |
| JayF | I'll take a look at those patches and re-review them, it'll be exciting to get those landed \o/ thank you for the db work | 21:04 |
| TheJulia | I needed to do soemthing that was not look at inspector today :) | 21:07 |
| opendevreview | Merged openstack/ironic stable/wallaby: Stable only: Factor out addition of packaging lib https://review.opendev.org/c/openstack/ironic/+/860144 | 21:52 |
| opendevreview | Verification of a change to openstack/ironic stable/xena failed: Stable only: Factor out addition of packaging lib https://review.opendev.org/c/openstack/ironic/+/860142 | 22:01 |
| JayF | ironic-standalone job has been failing on a large number of patches recently | 23:01 |
| JayF | just saying that out loud here if someone wants to look into it; there are some examples in the SQLA line ^^ up there, or really any random patch | 23:01 |
| JayF | otherwise I'll add to my list and see if I can get there tomorrow | 23:02 |
| opendevreview | Merged openstack/metalsmith stable/zed: Allow to use Ansible OpenStack Col. 1.x.x with openstacksdk >=0.99.0 https://review.opendev.org/c/openstack/metalsmith/+/861050 | 23:20 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
