Tuesday, 2025-11-18

tonybthe RH folks that care about os-net-config tried to continue with hosting it on opendev but the appetite seemed low so they went elsewhere 07:29
tonybAlos mnasiadka meeting in roughly 30mins right?07:30
mnasiadkaAh right07:31
mnasiadkatc-members - meeting in ~28 minutes07:32
tonybmnasiadka: I didn't mean to nag, more confirming I have the correct date.07:35
mnasiadkatonyb: it’s correct - but it’s 8:35 AM - so pardon my non-focus :)07:35
tonyblol of course07:36
mnasiadka#startmeeting tc08:00
opendevmeetMeeting started Tue Nov 18 08:00:16 2025 UTC and is due to finish in 60 minutes.  The chair is mnasiadka. Information about MeetBot at http://wiki.debian.org/MeetBot.08:00
opendevmeetUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.08:00
opendevmeetThe meeting name has been set to 'tc'08:00
mnasiadkaWelcome to the weekly meeting of the OpenStack Technical Committee. A reminder that this meeting is held under the OpenInfra Code of Conduct available at https://openinfra.dev/legal/code-of-conduct.08:00
mnasiadkaToday's meeting agenda can be found at https://wiki.openstack.org/wiki/Meetings/TechnicalCommittee08:00
mnasiadka#topic Roll Call08:00
tonybo/08:00
mnasiadkao/08:00
gtemao/08:00
mnasiadkacourtesy-ping: gouthamr, noonedeadpunk, frickler, spotz[m], cardoe, bauzas08:02
mnasiadka#topic Last Week's AIs08:04
mnasiadkaAnalyze user survey results08:04
mnasiadkaWe need a volunteer for this - if there are none - it will get moved to TC tracker (in the hope somebody wants to pick it up later)08:05
mnasiadkaAssuming no volunteers - will move it to the tracker after the meeting08:06
gtemaI can't see for myself  what can be useful in that data in that form08:06
tonybI'm curious what analysis there is for us to do? but happy let it move to the  tracker08:06
mnasiadkaMe neither, but it seems usually somebody from the TC did an analysis and published that08:07
mnasiadka#link https://governance.openstack.org/tc/user_survey/analysis-2023.html08:07
mnasiadkaAs an example08:07
gtemait feels like a yet another survey where nobody has really idea what is this good for08:07
gtemaslaweq was preparing this overview if I recall correctly08:08
tonybI can see it's value for the foundation and for the TC in very specific cases08:08
mnasiadkaI personally feel the user survey has too many possible fields to fill out or the input is a bit vague (e.g. you can choose ‘Ansible’ as your deployment tool)08:08
gtemathe results analysis should be done by the people who came with those questions08:09
mnasiadkaOh it surely should08:09
mnasiadkaBut let’s move that to the tracker for now08:09
mnasiadkaOur discussion doesn’t really help in finding a solution for the analysis08:09
mnasiadkaNext one - Draft TC resolution on os-net-config handover (gouthamr)08:09
mnasiadka#link https://review.opendev.org/c/openstack/governance/+/967463 (2025-11-17 Retirement and Handover of the `os-net-config` Project)08:09
tonybI'll add my views to the review but in general I don't think there's alot for the 3 of us on that one?08:11
gtemaindeed08:11
mnasiadkaYeah, I just commented on this08:11
mnasiadkaLet’s move forward08:11
mnasiadkaPropose document change to redirect os-net-config docs link to git repo README (gouthamr)08:12
mnasiadkaNo need to discuss that08:12
mnasiadkaOk, next one is08:12
mnasiadkaRework FIPS goal objectives and testing requirements (volunteer needed, move to tracker if no-one interested)08:12
mnasiadkaI personally have no interest in FIPS, given it’s for one country - is there any volunteer interested in FIPS? If no - we’ll move that to the tracker (given no traction/interest to date)08:13
gtemaoh lol - move everything to tracker ;-)08:13
tonybget that ^^ man a beer/coffee!08:14
mnasiadkaAnd then we send the tracker to /dev/null? lol08:14
tonyb:D08:14
gtemacorrect - that was a not-expressed thought08:14
mnasiadkaWell, let’s keep it recorded - but if there’s no interest - we can’t really do anything08:14
mnasiadkaNext one08:15
mnasiadkaUpdate/document procedure to preserve project appointment history (tonyb)08:15
mnasiadkatonyb: any updates?08:15
* tonyb has no recollection of this item08:15
mnasiadkaHmm, that’s interesting - I think it came up regarding the recent telemetry PTL switch08:16
tonybCan we add a meta-AI for me to figure out what it is and then either do it or correct the AI08:16
tonybSounds plausible 08:16
mnasiadkaNo other option - I’ll leave it in AIs on the wiki :)08:17
tonyb\o/08:17
mnasiadkaNext one is: Clarify and document AI/LLM policy/instructions for Project Team Guide08:17
mnasiadkaBut cardoe (who’s name is against the AI in the wiki) is not here08:17
mnasiadkaSo let’s move on08:17
mnasiadkaInvestigate adding ""archived"" state to retired PyPI projects (volunteer needed, move to tracker if no-one interested)08:18
tonybI can look at that.  I'm not 100% certain of the value but I can look at the technical aspects08:18
mnasiadka#link https://blog.pypi.org/posts/2025-01-30-archival/08:18
mnasiadkaThis looks like something new this year in pypi08:19
gtemaI personally like very much when projects on github are explicitly marked archived 08:19
mnasiadkatonyb: thanks08:19
gtemait gives a clear signal - keep away08:19
mnasiadkaI’m not sure how different is retired vs archived in pypi08:19
mnasiadkaBut that’s probably for tonyb to find out :)08:20
gtemayeah08:20
tonybHopefully is adds a banner like "this project is read-only" similar to github and others08:20
mnasiadkaThat would be nice, I agree08:20
mnasiadkaOk, that’s all of the AIs from last week08:20
mnasiadka#topic Monasca Retirement08:20
mnasiadka#link https://review.opendev.org/q/hashtag:%22monasca-retirement%22+(status:open%20OR%20status:merged)08:21
mnasiadkaThe repository cleanup patches need to be merged. However, openstack/monasca-api and openstack/monasca-events-api are failing with zuul errors.08:21
mnasiadkaIt's possible older stable and unmaintained branches need to be deleted too, to fix the cleanup on master. Need some help from the release team or TaCT SIG on how to proceed08:21
mnasiadkaThat’s status from gouthamr08:21
mnasiadkatonyb: It seems that the foundation member you mentioned hasn’t really acted - do you have any more information on that?08:22
tonybgive me the AI I can look at it from a TaCT POV08:22
gtemait upsets me how a single item can hold the whole community for 4-5 month now already? Just kill it and let people move forward, why are we taking so long on making an action08:22
gtemasorry, this was rhetoric08:23
mnasiadkagtema: I agree the time to do it was long time ago, so let’s at least not stall it this time - I’ll try to help gouthamr in progressing that08:23
mnasiadkatonyb: AI granted :)08:23
tonybmnasiadka: It looks like *if* they're going to use it it wont be for 12months08:23
mnasiadkaAh, *if*, then I don’t think we should really still wait.08:23
mnasiadkaOk then, going forward08:24
mnasiadka#topic TC Tracker, PTG Follow up08:24
mnasiadka#link https://etherpad.opendev.org/p/tc-2026.1-tracker (Technical Committee activity tracker - 2026.1)08:24
tonybYup I thought we'd agreed to retire monasca I'm sorry if I was holding things up by mistake08:25
noonedeadpunkoops08:25
* noonedeadpunk completely forgot about morning one08:25
mnasiadkaThere’s the OpenStack Survey improvements topic there - I feel we should try to work for the data (that usually come from manual analysis) to be visible in a more dynamic way (maybe on the user survey webpage or something similar)08:25
mnasiadkaThe deadline for submitting questions to survey is 1st Dec 2025 - for these that are interested08:26
mnasiadkaThere’s also an etherpad with TC PTG summary08:27
mnasiadka#link https://etherpad.opendev.org/p/oct2025-ptg-os-tc-summary08:27
mnasiadkaAnd a mail by gouthamr to the openstack-discuss ML08:27
mnasiadka#link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/message/RX3MZE33GEDV5JDHORZKRUKVDP47UMLJ/08:27
mnasiadkatonyb: looking at tracker L119 and L124 - seems you have some AI in there - although I don’t know what needs to be done08:28
tonybmnasiadka: I do, I just haven't done it08:28
mnasiadkanoonedeadpunk: no worries, no AIs on you (yet) ;-)08:29
mnasiadkatonyb: thank you for the update :)08:29
tonybLOL08:29
mnasiadkaOk then, to next topic08:29
mnasiadka#topic Python PTI and test runner guidance08:29
mnasiadka#link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/3V3CNPQLB77SKFVLZ6LXJ5NPNYWW4QFD/08:29
gtemamdasiadka and noonedeadpunk: wrt PQC, I am supporting that to a certain extend and wanted to drop passlib (a completely dropped lib) from our requirements and noticed that it is now only used in the deployment toolings08:29
mnasiadka(Request for guidance on improving Python PTI doc to include pytest for Horizon plugins testing)08:29
gtemawould you be open still getting rid of it? Well, it is not a question honestly - this must be done. Period.08:30
mnasiadkagtema: Sure, happy to do it in Kolla-Ansible - can you maybe start an openstack-discuss ML thread pointing to the repositories using it and we could track it that way?08:31
gtemahttps://opendev.org/openstack/-/code?q=passlib&search_mode=exact08:31
mnasiadka#link https://codesearch.opendev.org/?q=passlib&i=nope&literal=nope&files=&excludeFiles=&repos=08:31
gtemabut sure, I can start the thread and prepare change dropping it from requrements08:32
noonedeadpunkgtema: looking at codesearch, osa does not need it08:32
gtemaright, but who is openstack/openstack-ansible-ops?08:32
noonedeadpunkbut bifrost on other hand seems tro rely on it?08:33
tonybgtema: I think that the openstack-ansible SIG?08:33
gtemaand who nowadays owns openstack/deb-* repos08:33
noonedeadpunkyeah, osa. but you can drop it and let it fail :)08:33
gtemaah, ok08:33
noonedeadpunkor well. I can clean it up later this week08:33
gtemathanks.08:34
noonedeadpunkI think openstack-ansible-ops is not actually following requirements framework anyway08:34
gtemaSorry for the delayed message that broke the topic boundary08:34
noonedeadpunkit's pile of operator tooling, half of which should be refactored or dropped at this point anyway08:34
mnasiadkaK-A only uses passlib[bcrypt] for generating passwords, we can surely use something else08:35
mnasiadkagtema: any recommendations? ;-)08:35
noonedeadpunkwe use haslib I think08:35
gtemabcrypt is what keystone uses by default, but today the modern recommendation is to use argon208:36
noonedeadpunknot sure if it's better 08:36
gtemahashlib is bad, you should be using salt based hashing. In germany this is a requirement on the governmental level08:36
gtema#link https://argon2-cffi.readthedocs.io/en/stable/08:37
noonedeadpunksalt hashing for random password generator?08:37
gtemathis is what people should be using today08:37
mnasiadkaOk, I’ll have a look in argon208:37
gtemawhenever you store password hash in the DB it must be salt based hashing08:37
noonedeadpunkcan you supply a salted password to keystone when creating a service user?08:37
gtemanot the salted password - the salted hash of the password08:38
noonedeadpunkbut how you'd call openstack.cloud.identity_user with that?08:39
gtemabcrypt is still safe, but argon is more future proof08:39
noonedeadpunkor, `openstack user create`?08:39
gtemanoonedeadpunk - nothing changes on the password logic itself, only when you store the hash in the DB you should be applying salt to it - it has nothing to do with ansible modules for openstack08:40
noonedeadpunkyes, but there's no database. And for storage it's supposed that Ansible Vault will be used08:40
noonedeadpunkBut you need to generate passwords in more or less plain text so that they could be supplied to services to create db/keystone/rabbitmq/etc users08:41
noonedeadpunkso in context of deployment tools salting is irrelevant I'd assume08:41
tonyb... Can we move this to after the meeting.  It's valuable but also a tangent.08:41
noonedeadpunk++08:41
gtemahere there is nothing what we can do so far - those passwords stay in plain text in config files08:41
mnasiadkaOk, let’s move it after meeting08:42
mnasiadkaAnd get back to the original topic - pytest for Horizon plugins08:42
noonedeadpunkmnasiadka: fwiw https://opendev.org/openstack/openstack-ansible/src/branch/master/scripts/pw-token-gen.py - the thing we have08:42
gtemawrt pti I would like to recall the discuss thread of Clark - we should establish a platform for reviewing old rules08:42
gtemaI think I have expressed my opinion in multiple threads clear enough08:43
* noonedeadpunk need to get time to finally respond to these threads08:43
gtemaand the horizon folks themselves made a summary of the discussions to clarify their next steps08:44
mnasiadkaI agree with gtema - but I think this needs more people than 4 on a TC meeting - especially that noonedeadpunk is not up to date with the thread08:45
gtemasure08:45
noonedeadpunkI read it, I just never managed to respond as so many things to cover....08:45
mnasiadkaOk, I’ll leave that for the next meeting08:45
mnasiadkaTo the next topic...08:45
mnasiadka#topic oslo.wsgi08:45
mnasiadka#link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/IW2ZMYXGZQFPSLJRVT5OKFFX7XRGM2FF/08:46
gtemanot sure what is expected to be discussed here08:46
mnasiadkaTL;DR - there’s a proposal to create oslo.wsgi - all of the information is in the thread08:46
gtemago for it, but do not make it a MUST08:46
noonedeadpunkI think there was a general agreement that it's a good idea?08:46
tonybI feel like that's squarely in the oslo teams perview08:46
mnasiadkatonyb: ++08:47
tonybWe can reply with "match the import and governance patches" which is all I can see as being missing?08:48
mnasiadkaSo I guess there’s nothing to discuss, the thread replies from Oslo team and others indicate it’s a good idea - so I assume the requestor will follow up with governance patches08:48
mnasiadkaMoving forward08:48
mnasiadka#topic A check on gate health08:48
mnasiadkaIs there anything worth mentioning?08:48
tonybI can't think of anything.08:49
tonybgerrit and gitea have been updated in the last week 08:49
tonybbut they both went well08:49
mnasiadkaOk then, happy to hear that08:50
mnasiadka#topic Open Discussion and Reviews08:50
tonybWe (infra-root) "discovered" a cool new zuul feature in the process08:50
*** elodilles_pto is now known as elodilles08:51
tonybI don't have anything for Open Discussion08:51
gtemawhich is it?08:51
mnasiadkatonyb: cool? Which is?08:51
tonybgtema: You can pause the scheduler, so that if gerrit is unavailable it wont trigger a merge failure and then force those commits to be re-tested08:52
noonedeadpunkoh, nice08:52
tonybIt helps in a very narrow window but when it does it saves hours of "recheck"08:52
gtemaI haven't had merge failures often. It is usually a regular timeouts recheck madness08:53
mnasiadkatonyb: ah, that’s what happened with my patch when it was “ready to submit” during some Zuul restart window ;-)08:53
noonedeadpunkwhen you restart gerrit or upgrade gerrit - that really helpful indeed08:53
tonybYeah this is really for when gerrit is down/unavailable08:53
tonybmnasiadka: That's different but similar08:54
mnasiadkaack08:54
mnasiadkaOk then, I see no open discussion topics08:54
mnasiadkaThank you all for attending, see you next week in the regular timeslot08:55
mnasiadka#endmeeting08:55
opendevmeetMeeting ended Tue Nov 18 08:55:04 2025 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)08:55
opendevmeetMinutes:        https://meetings.opendev.org/meetings/tc/2025/tc.2025-11-18-08.00.html08:55
opendevmeetMinutes (text): https://meetings.opendev.org/meetings/tc/2025/tc.2025-11-18-08.00.txt08:55
opendevmeetLog:            https://meetings.opendev.org/meetings/tc/2025/tc.2025-11-18-08.00.log.html08:55
tonybThanks all08:55
gtemathanks, see ya08:55
mnasiadkanoonedeadpunk: Now we can continue the keystone/salts/passwords/whatever topic ;-)08:55
gtemasure, but in 5 min I need to jump to next meeting08:55
gtemabut is then available 20 min later08:55
gtemas/is/am/08:56
noonedeadpunkyeah. I think we were just talking about different things09:01
noonedeadpunkas there's nothing to store for us in salted manner, as there's no db for osa09:01
noonedeadpunkthe db is vars files containing secrets that are to be written to service dbs09:01
gtemayes, I think so. Unless you store passwords somewhere on the FS to be able to validate whatever user passes (like the /etc/passwd) - you are good09:13
fricklermeh, I totally missed that we have the early time slot this week, did take extra care to reserve the evening even if I'm mostly not around09:56
fricklerregarding monasca, I did plan to look at the retirement patches and force-merge them if necessary. only noticed last week that they were un-wip-ed09:57
fricklerfor FIPS, if there is no interest in continuing with the goal, maybe someone has interest to do the opposite and mark the goal as terminated? and possibly clean up broken jobs?09:58
bauzasoh shit, missed the TC meeting :(10:48
bauzasTheLounge can provide me some notifications but if I don't see them, I forget them :(10:49
mnasiadkaThat’s completely fine everybody from EU is complaining about regular meeting time, but the one time in the month we all forget about it (and I remembered about it only because gouthamr reminded me about it) :)10:58
bauzasI'll just add the ICS file in my google agenda11:00
bauzasjust in case people don't know : https://meetings.opendev.org/#Technical_Committee_Meeting11:01
cardoeYeah I’m not gonna make the EU timed ones. It’s 2am to 3am for me.13:39
fungijust a quick follow-up on the password hashing discussion, salted (and peppered) hashes are still an outdated notion. use a key derivation function (kdf) with adjustable multi-round options and padding and such15:03
fungi(algorithms like bcrypt and argon2 are kdfs, so that's fine, it's more a matter of terminology)15:05
clarkbya I was kind of hoping for more engagement on whether or not we think it is a good idea to formalize the 'evolution' process more concretely16:02
clarkbI don't think the actual mechanics of the process matter, but do believe performing some sort of sanity checking and updates as necessary is likely to be important16:02
fungialso, in the spirit of chesterton's fence, make sure any rule you're seeking to get rid of or replace is thoroughly understood first. don't take something away until you are clear on why it was implemented and what benefit it provided/still provides16:05
fungiopenstack has a tradition of avoiding the addition of red tape and bureaucracy unless there's sufficient solid benefit to more than offset it, so looking at the history of a particular decision and getting a grasp of why it was made at the time is necessary if you want to avoid rediscovering problems we already solved16:07
mnasiadkaI think that discussion needs to wait for next Tue - to have a proper quorum on the meeting…16:17
clarkback. Its a holiday week in the US (particularly thursday and friday), but I should be around tuesday and can discuss further then16:18
clarkbholiday week next week; not this week16:18
mnasiadkaI’m not from the OpenStack historian society - but will the evolution require us to dig out old decisions that are not centrally stored?16:27
clarkbI think right now its mostly just a prompt to talk through these concerns. My personal perspective is that I think we as a community need to be more self deterministic and make decisions that work for the community arther than letting things happen and being forced to make changes.16:28
clarkbI think in general a lot of decisions have been reasonably well recorded (the historical data largely being in TC documents in git and gerrit for example)16:28
fungimnasiadka: my usual tools for researching these decisions is git and gerrit, looking at the discussions in gerrit changes that implemented those choices and subsequent references to sources like tc meeting logs or mailing list posts16:31
mnasiadkaAck, just trying to understand if we have a list of decisions in mind16:32
clarkbfor me the big ones that have come up recently are the PTI (and which platforms we map to the PTI eg centos stream) and which languages we accept. Golang was added and today we have zero code written in golang that we maintain aiui (so less even than a qusetion about rust, but its clear golang hasn't been effective imo). But also things like translation tooling (whcih is16:34
clarkbcurrently being actively modified but again due to forced decision making)16:34
fungibut basically anything documented in the governance repo is fair game to revisit if people feel something there is getting in the way16:35
fungithe pytest vs unittest decision in the pti falls into that category too based on recent discussion about horizon maintainers disagreeing with the that expectation16:36
clarkband we can go further into things like wsgi which in our world really became uwsgi. There are multiple alternatives to wsgi now with a whole host of options in terms of server choice too16:37
clarkbeventlet was adopted far and wide, though I'm not sure if this was ever really official beyond the decision to update nova from twisted16:37
clarkbI think there is value in consistency for developers, operators, and users. But there is also value in being able to use appropriate tools for specific problems. That balance is likely to shift over time so I'm suggesting that we think about this balancing act more explicitly rather than only in reaction to emergencies16:40
spotz[m]o/17:06
spotz[m]Well poo I had it on my calendar starting now!17:06
fungiyou probably wouldn't have wanted to join a 2am meeting anyway17:11
clarkbI definitely didn't17:17
fungiwould have been midnight for you, i think17:20
clarkbPeter Coyote had me falling asleep just after 10pm17:22
* gouthamr catches up18:32
gouthamrthank you for running the meeting mnasiadka \o/ 18:32
gouthamrfrickler: thanks for offering help with the 'problematic' monasca repos18:33
gouthamri do like that tonyb made it, but hopefully, you folks can too.. remember though, going into the end of the year, a few meetings may be canceled. 18:33
gouthamryou folks: for anyone that can reasonably get to their computer without having to douse themselves in caffiene 18:34
fungii bathe in the stuff, but 3am is a struggle... too late to not be too late, while too early to not be too early18:35
gouthamr:D 18:35
fungii can mostly convince myself to wake up at 4am now and then, but intentionally staying awake until 3am doesn't really happen at my age18:36
fungiit's insane to think back to the days of my youth when i'd go out clubbing until they kicked us out, then find 3am breakfast at an all-night diner, and get home with just enough time to shower and stagger into the office18:38
* gouthamr feels like a long time ago, but didn't 19:02
gouthamri was a regular at Cookout :P 19:02
gouthamromg fungi, if this is any close: https://www.outerbanksvoice.com/2025/11/08/cook-out-opens-its-doors-in-kdh/ 19:05
fungiyeah, they had a line around the block19:05
fungithat's only a few miles from my house19:06
gouthamrhahaha, that's true for every day at 3am in Clemson, and probably lots of other places19:06
gouthamrvery nice! 19:07
fungii don't know how long it'll survive here though, the local community isn't very fond of national chains and the 10x population swing between tourist season and off-season means restaurant businesses need to be able to adapt and staff up and down seasonally19:07
clarkbhere in portland we had Le Bistro Montage. They are only a food cart now but once upon a time they had a restaurant under a bridge in the industrial area and were open all night19:07
gouthamralways need for that, we sleep too early19:09
opendevreviewGoutham Pacha Ravi proposed openstack/openstack-manuals master: Re-direct tripleo docs to repos  https://review.opendev.org/c/openstack/openstack-manuals/+/96761722:06
opendevreviewGoutham Pacha Ravi proposed openstack/governance master: 2025-11-17 Retirement and Handover of the `os-net-config` Project  https://review.opendev.org/c/openstack/governance/+/96746322:45

Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!